The fortunes of Byzantine Constantinople have always been inextricably linked to the sea.

The
The Byzantine Harbours

The Byzantine Harbours of Constantinople

­topographical, demographic and economic development of the city and its networks are reflected in

of Constantinople
the history of its harbours. This volume offers an exhaustive study of Constantinople’s Byzantine harbours
on the Sea of Marmara and the Golden Horn, as well as nearby European and Asian landing stages.
The fifteen chapters by eleven contributors here present a broad synthesis of the current state of research
using written, pictorial and archaeological sources.

Falko Daim · Ewald Kislinger (eds)

Daim · Kislinger (eds)

Byzanz zwischen Orient und Okzident:

Veröffentlichungen des Leibniz-WissenschaftsCampus Mainz / Frankfurt

Die Reihe Byzanz zwischen Orient und Okzident wird vom Vorstand des gleichnamigen
Leibniz-WissenschaftsCampus, einer Kooperation des Römisch-Germanischen Zentralmuseums,
der Johannes Gutenberg-Universität Mainz, der Goethe-Universität Frankfurt und des Leibniz-Instituts
BOO  Band 24

für Europäische Geschichte in Mainz, herausgegeben.

Die Reihe dient als Publikationsorgan für das Forschungsprogramm des Leibniz-WissenschaftsCampus, das
Byzanz, seine Brückenfunktion zwischen Ost und West sowie kulturelle Transfer- und Rezeptionsprozesse
von der Antike bis in die Neuzeit in den Blick nimmt. Die Methoden und Untersuchungsgegenstände der
verschiedenen Disziplinen, die sich mit Byzanz beschäftigen, werden dabei jenseits traditioneller Fächer-
grenzen zusammengeführt, um mit einem historisch-kulturwissenschaftlichen Zugang Byzanz und seine
materielle und immaterielle Kultur umfassend zu erforschen.
Byzanz zwischen Orient und Okzident  Veröffentlichungen des Leibniz-WissenschaftsCampus Mainz / Frankfurt
 Byzanz zwischen Orient und Okzident  | 24
Veröffentlichungen des Leibniz-WissenschaftsCampus Mainz / Frankfurt

zugleich

Interdisziplinäre Forschungen
zu den Häfen von der Römischen Kaiserzeit bis zum Mittelalter in Europa
C. von Carnap-Bornheim  ·  F. Daim  ·  P. Ettel  ·  U. Warnke (Hrsg.)

Band 10

RG Z M
The Byzantine Harbours
of Constantinople

Falko Daim  ·  Ewald Kislinger (eds)

Verlag des Römisch-Germanischen Zentralmuseums | Mainz | 2021

Redaktion: Stefan Albrecht (RGZM)
Satz: Claudia Nickel (RGZM)
Cover: Dominik Heher, Claudia Nickel (RGZM),
unter Verwendung eines Bildes von Antoine Helbert
Übersetzung: Leo Ruickbie, Antje Bosselmann-Ruickbie

Bibliografische Information
der Deutschen Nationalbibliothek

Die Deutsche Nationalbibliothek verzeichnet diese Publikation in

der Deutschen Nationalbibliografie; detaillierte bibliografische
Daten sind im Internet über http://dnb.d-nb.de abrufbar.

ISSN 2626-9392 (Print)

ISSN 2629-2769 (Online)
ISBN 978-3-88467-344-7

Dieses Werk ist unter der Creative Commons-

Lizenz 4.0 (CC BY-SA 4.0) veröffentlicht.

Diese Publikation ist auf http://www.propylaeum.de

dauerhaft frei verfügbar (Open Access).
DOI: https://doi.org/10.11588/propylaeum.911
eISBN: 978-3-96929-086-6
URN: urn:nbn:de:bsz:16-propylaeum-ebook-911-0

© 2021 Verlag des Römisch-Germanischen Zentralmuseums

Das Werk ist urheberrechtlich geschützt. Die dadurch begründeten

Rechte, insbesondere die der Übersetzung, des Nachdrucks, der
Entnahme von Abbildungen, der Funk- und Fernsehsendung, der
Wiedergabe auf fotomechanischem (Fotokopie, Mikrokopie) oder
ähnlichem Wege und der Speicherung in Datenverarbeitungs-
anlagen, Ton- und Bildträgern bleiben, auch bei nur auszugsweiser
Verwertung, vorbehalten. Die Vergütungsansprüche des
§ 54, Abs. 2, UrhG. werden durch die Verwertungsgesellschaft
Wort wahrgenommen.

Druck: Beltz Grafische Betriebe GmbH, Bad Langensalza

Printed in Germany.
Contents

Falko Daim · Ewald Kislinger

  7 Foreword

Ewald Kislinger
  9 On Better and Worse Sites: The Changing Importance of the Harbours of Constantinople

Arne Effenberger
 19 Constantinople / İstanbul: The Early Pictorial Sources

Alkiviadis Alexandros Ginalis · Ayşe Ercan-Kydonakis

Some Reflections on the Archaeology of the Late Antique and Byzantine Harbours
 33 of Constantinople

Harbours and Landing Stages of Constantinople

Andreas Külzer
 75 The Harbour of Theodosius in Yenikapı, İstanbul: A Harbour Area through the Ages

Dominik Heher
 93 Harbour of Julian – Harbour of Sophia – Kontoskalion

Dominik Heher
109 The Harbour of the Bukoleon Palace

Ewald Kislinger
133 Neorion and Prosphorion: The Old Harbours on the Golden Horn

Johannes Preiser-Kapeller
141 Heptaskalon and Other Landing Stages on the Golden Horn

Peter Schreiner
The Western Landing Stages (σκάλαι) in the Golden Horn:
151 Some Remarks Relating to the Eleventh and Twelfth Centuries

Neslihan Asutay-Effenberger
161 The Kynegion District and its Harbour in Late Byzantine and Ottoman Times

Ewald Kislinger
171 The Golden Horn: Constantinople’s Superharbour (Überhafen) and its Chain

Harbours, Landing Stages, and Moorings of Constantinople’s Outskirts

Grigori Simeonov
181 The Harbour and Landing Stages of Hebdomon

Grigori Simeonov
199 The Landing Stage of Brachialion
 Grigori Simeonov
209 The Moorings at Kosmidion

Klaus Belke
Gates to Asia Minor: The Harbours of Chalcedon, Chrysopolis,
223 Hiereia and Eutropiu Limen opposite Constantinople

235 Maps of Constantinople

239 Bibliography

273 Index (compiled by Klaus Belke)

285 List of Contributors

287 Sigles Used

Translators‘ note:
Each author has made an individual decision on the style of naming places and people,
including whether to use Latin or Greek forms, meaning that spellings will vary throughout
the book, which is reflected in the index.
 Foreword
During the work on the Istanbul underground about twenty
years ago, the remains of the medieval Harbour of Theodosius
were rediscovered, along with nearly 40 shipwrecks. This was
an archaeological find of the century and a substantial argu-
ment for including the Byzantine harbour landscape in the
Priority Programme 1630 of the DFG (Deutsche Forschungs-
gemeinschaft / German Research Council) »Harbours from the
Roman Imperial Period to the Middle Ages«.
One of the individual projects within this Priority Pro-
gramme is »Ports and Landing Places on the Balkan Coasts of
the Byzantine Empire (Fourth to Twelfth Century): Technology
and Monuments, Economy and Communication«. It is part
of the Leibniz ScienceCampus Mainz / Frankfurt: Byzantium
between Orient and Occident, a collaboration between the
Römisch-Germanisches Zentralmuseum in Mainz (RGZM), the
Johannes Gutenberg University Mainz and other partners, es-
pecially the Viennese School of Byzantine Studies (at the Uni-
versity of Vienna and Austrian Academy of Sciences / ÖAW).
As was generally the case throughout history, Constan-
tinople also played a central role in seafaring. Situated at the
southern mouth of the Bosphorus on a peninsula between
the Sea of Marmara and the Golden Horn, the daily life of the
inhabitants in many respects depended on the sea, and the
harbours formed the interface between the city and the sea.
For centuries, Constantinople was one of the most important
and powerful trading centres in the Mediterranean. In addi-
tion, the Byzantine Empire also dominated the Mediterranean
militarily for a long time (thalassocracy) and its main fleet was
stationed in Constantinople.
Written sources mention a number of smaller and larger
harbours that were repeatedly rebuilt, renamed or even
newly built over the centuries. These harbours have been
researched over the years. Wolfgang Müller-Wiener earlier
brought these results together in his work »Die Häfen von
Byzantion – Konstantinupolis – Istanbul« (»The Harbours
of Byzantion, Constantinople, Istanbul«), published posthu-
mously in 1994. However, research on the topography of the
city and on Byzantine seafaring, which has intensified since
then, has yielded new data and perspectives.
During a special course held at the University of Vienna
in 2014, the idea arose to not only include the harbours of
Constantinople in the DFG project’s catalogue in an overview,
but also dedicate a separate anthology to them that would
In: Falko Daim  ·  Ewald Kislinger (eds), The Byzantine Harbours of Constantinople. Byzanz zwischen Orient und Okzident 24 (Mainz 2021).
DOI: https://doi.org/10.11588/propylaeum.911.c12062
reflect the current state of research. The development of
the Constantinopolitan harbours covers the entire Byzantine
period from the fourth to the fifteenth centuries. In addition,
the immediate periphery must also be considered: on the
one hand, the harbours on the Asian side of the Bosphorus;
and, on the other hand, the landing stages in the immediate
Thracian vicinity of the city walls, both on the Golden Horn
and on the Sea of Marmara.
The first edition of this book was published in 2016 in
German as Volume 4 of the Leibniz ScienceCampus’s book
series »Byzantium between Orient and Occident« (BOO). The
eight contributors produced a total of twelve essays and have
all worked within the frame of the Priority Programme 1630
of the DFG or co-operated with it. This was reviewed several
times, mostly appreciatively, but also with suggestions for
possible deepening and broadening of the contents.
In response to this, the two editors of this volume, Falko
Daim and Ewald Kislinger, decided in 2020 to publish an ex-
panded version in English to reach a broader audience. The
original contributions were not only translated but updated,
and the now eleven authors of different nationalities and
mother tongues have delivered a total of fifteen essays for
the English edition. They have endeavoured to reflect the
broad linguistic spectrum of publications on the topic in
Byzantine studies, considering a narrow approach (as seen
in some recent companions) to be less fruitful. All publica-
tions on the subject that were published, known to us, and
accessible by the end of 2020 have been taken into account.
Without the commitment of all the authors, this volume
could not have been completed so swiftly, including the
addition of an index, which was the special responsibility
of Klaus Belke. Johannes Preiser-Kapeller and Michael Ober
were in charge of the maps and plans. Our sincere thanks
must also go to other colleagues: Leo Ruickbie and Antje Bos-
selmann-Ruickbie, who prepared the English translations with
great dedication and expertise, with additional input from
Laury Sarti, and provided editorial co-ordination, and Franz
Siegmeth, who prepared various illustrations for printing. We
would also like to express our thanks to Claudia Nickel and
Stefan Albrecht at the RGZM publishing house.
As has always been the function of the harbours them-
selves, this volume is not intended to be an end but a starting
point for new research.
Mainz and Vienna, March 2021
Falko Daim and Ewald Kislinger
Ewald Kislinger
On Better and Worse Sites:
The Changing Importance
of the Harbours of Constantinople
The harbours of ancient Byzantion and later Constantinople
in its early days were orientated towards the south-eastern
coast of the Golden Horn where it opens into the Bosphorus.
Neorion and Prosphorion 1 were situated here, semi-circular
harbours of the old type, probably lined with colonnades, and
also the ferry crossings to Chalcedon und Sykai (Pera) 2. The
Harbour of Julian was added in the south of the peninsula,
facing the Sea of Marmara, soon after Constantinian expan-
sion of Constantinople 3. The original name of this harbour
provides an approximate dating, especially since there was
a statue of Emperor Julian (reg. 360-363) 4. It collapsed in
the year 533, which, together with the construction work of
Justin II (reg. 565-578) 5, paved the way for a renaming after
his wife Sophia. Further to the west was the small Harbour of
1 On the two harbours, see my chapter on Neorion and Prosphorion in this vol-
ume  – Cf. Janin, Con­stan­tinople 236  – Müller-Wiener, Häfen 6-7.  – Berger,
Regionen und Straßen 362. 364-365. – Schreiner, Costantinopoli 106-107. –
Leszka, Konstantynopolitańskie porty 6. 8. 12. – Magdalino, Harbors 13-14 (un-
documented); Stavroulaki, Seaports (without merit). – Dark, Eastern Harbours
152-163 (cf. Dark, New Post Office). – Berger, Häfen von Byzanz und Konstantin-
opel is for the better part based (as admitted by the author, 111) on the research
by Müller-Wiener, Häfen, therefore, does not need to be quoted here. Berger,
Häfen 77. 80-81 contains a number of borrowings from the above-mentioned
article, but with additional (though sparse) footnotes. – Günsenin, »City« Har-
bours 99-105 is based (see 99) on Janin, Constantinople, Müller-Wiener, Häfen
and Mango, Développement, Recent archaeological evidence is considered in
the short overview, particularly on the district of Sirkeci. A number of mistakes
in writing of historical denominations are striking: Portus Theodasiacus and Kon-
taskalion (op. cit. 103. 104 with n. 2). Improved in Günsenin, Harbours and
Shipbuilding 414-416.
2 Notitia urbis Con­stan­tinopolitanae 233, 15 (scala Timasii); 234, 19 and 11. –
Chronicon Paschale 569 and 572 (Dindorf): the relics of three saints (Samuel,
Joseph, the son of Jacob, and Zachariah) landed in 406 or 415 at the skala
Chalkedonensia. On their location, see Berger, Regionen und Straßen 362. 364. –
For completeness, the landing place below the Arcadianai baths in the Acropolis
area is mentioned: Prokopios, De aedificiis I 11, 1-2 (IV 41 Haury / Wirth), where –
obviously, so as not to cause a stir – the imprisoned Pope Martin I was disem-
barked before his trial in 653, according to his Greek vita (Vita Martini ch. 6 [258
Peeters]). See also, Chiesa, Biografie 216 n. 10. – Also, the Mangana, the arsenal
for weapons and siege equipment, would have had access to the sea, as well as
the south-lying palace of the same name (Schneider, Mauern und Tore 95 and
105 [plan 5]. – Demangel / Mamboury, Quartier des Manganes 7-8 n. 2 pl. I-II).
3 See the contribution by Heher, Harbour of Julian, in this volume. – Cf. Mül­ler-
Wiener, Häfen 8-9. – Janin, Con­stan­tinople 231-234. – Mango, Développement
38-40. – Magdalino, Con­stan­tinople 20-22. – Berger, Regionen und Straßen
360-361. – Berger, Häfen 82-83. 85. – Leszka, Konstantynopolitańskie porty
7-9. 13. 15. – Günsenin, Harbours and Shipbuilding 417. – The Kontoskalion
Harbour cannot be equated with the Eleutherios Harbour (slightly east of the
Theodosios Harbour), as assumed by Günsenin, »City« Harbours 104 and Iva-
nov, Konstantinopol 416-418. The latter had been filled with the excavated
material from the former during the construction of the Forum Tauri under
Emperor Theodosius I (379-395), see Külzer, Harbour of Theodosius, in this
In: Falko Daim  ·  Ewald Kislinger (eds), The Byzantine Harbours of Constantinople. Byzanz zwischen Orient und Okzident 24 (Mainz 2021).
DOI: https://doi.org/10.11588/propylaeum.911.c12063
Eleutherios, said to have been created under Constantine the
Great. More details are unknown because it was filled in with
excavated earth from levelling work for the erection of the
Column of Theodosius in the Forum Tauri 6. It was generously
compensated for by the westward re-establishment of the
Harbour of Theodosius (probably equivalent to the Harbour
of Caesarius) at the mouth of the Lykos / Lycus 7.
The urban growth – whose two phases are clearly marked
by the Constantinian and Theodosian city walls 8  – and in
conjunction with that the expansion of the harbours, is thus
concentrated in the south, on the coast of the Propontis.
Until the seventh century, however, there was still a balance
between the infrastructure and the flow of goods, such as
food supplies, but also building materials like timber and
volume. – In the Vlanga quarter (in the area of the Theodosios Harbour, not at
the Kontoskalion, see Günsenin, »City« Harbours 104), Jewish tanners were
settled in the Palaeologan period (who before had been active near the south
banks of the Golden Horn, see below). The contributions by Berger, Langa
Bostanı (1993) seems to have been unknown to Günsenin, »City« Harbours
(2012). – On the private imperial harbour a little further east, see Heher, Bou-
koleonhafen, and Heher, Harbour of the Bukoleon, in this volume. – Berger,
Häfen 83 without knowledge of Heher, Bukoleonhafen. – Günsenin, »City«
Harbours 103.
4 Zosimos, Historia Nova III 11 (II 25 Paschoud). – Prokopios, De aedificiis I 4, 28 (IV
26 Haury / Wirth). – Ioannes Malalas, Chronographia XVIII 82 (404 Thurn). – The
Notitia urbis Con­stan­tinopolitanae 232, 9 referred to as the portus novus in re-
gion III. Berger, Regionen und Straßen 360-361; English translation in Matthews,
Notitia 88. – For the dating of the sources, see Berger, Regionen und Straßen
350-351: main part around 425.
5 Leon Grammatikos, Chronographia 135 (Bekker). – Symeon Logothetes, Chron-
icon 147 (Wahlgren). – Patria Kon­stan­tinupoleos III 37 (230 Preger). – Damage
caused by fire may have preceded it (Ioannes Malalas, Chronographia XVIII 131
[422 Thurn]). – Theophanes, Chronographia 235 [de Boor]).
6 Patria Kon­stan­tinupoleos II 63 and III 91 (184-185 and 248 Preger). – Müller-Wie-
ner, Häfen 9 n. 25: located below the Myrelaion complex. – Berger, Untersuch-
ungen 581-582. – Mango, Développement 55. – Cf. Külzer, Harbour of Theodo-
sius, in this volume.
7 Notitia urbis Con­stan­tinopolitanae 239: portus Theodosiacus in region XII. Mat-
thews, Notitia 95. – See Külzer, Harbour of Theodosius, in this volume. – Cf.
Müller-Wiener, Häfen 9. – Janin, Con­stan­tinople 226-228. – Mango, Dévelop-
pement 39-40. – Berger, Regionen und Straßen 372-373. – Kislinger, Lebens-
mittel. – Ercan, Yenikapı. – Leszka, Konstantynopolitańskie porty 10-11. – Güns-
enin, Harbours and Shipbuilding 417-418. However, the Belisar Tower was not
situated in the area of the Theodosios Harbour (see Günsenin, op. cit. 418), but
at the western end of the Boukoleon Harbours. – Pulak / Ingram / Jones, Yenikapı
102-103. – On the identification, see Guilland, Études de topographie II 95-96
and Berger, Untersuchungen 575. Like the Harbour of Julian, that of Theodosios
may previously have been a bay: Mango, Shoreline 20 fig. 1; Günsenin, Harbours
and Shipbuilding 419.
8 Asutay-Effenberger / Effenberger, Eski Imaret Camii.  – Asutay-Effenberger / Ef-
fenberger, Verlauf der Kon­stan­tinsmauer. – Asutay-Effenberger, Landmauer. –
Müller-Wiener, Bildlexikon 286-311 (each with older literature).
On Better and Worse Sites  |  Ewald Kislinger 9
bricks 9, which is reflected in the various granaries: the hor­
rea Troadensia, Valentiaca and Con­stan­tiaca are located in
the fifth region near the Strategion 10 (and thereby near the
Harbour of Prosphorion). Periodic inspections by the emperor
are testified only for these 11, but this could be due to the fact
that the custom originated in earlier times, when the later
horrea Alexandrina and the Theodosius granary in the ninth
Region on the Sea of Marmara (on the eastern edge of the
Harbour of Theodosius in region XII) 12 had not yet existed. At
the Harbour of Julian / Sophia, the Church of St Thecla en tois
krithopoleiois 13 indicates the handling and sale of barley and
another granary, which in addition to the five other (probably
larger) granaries mentioned in the Notitia urbis Con­stan­
tinopolitanae (242, 29 Seeck) signifying a balanced devel-
opment of such facilities on the Golden Horn and Propontis.
However, the period during which the different horrea
were active was varying. The granary complex at the Harbour
of Theodosius in the southwest was known as tēs Lamias
from the seventh century 14 (which incidentally confirms the
ongoing operation of the harbour even before the recent ship
finds 15), not far from the square tou Amastrianou 16. This in
turn was located just north of the harbour. According to a
 9 Themistios, Oratio 6. 83c-d and Or. 27, 336 d (124 Schenkl / Downey 160-161;
Downey / Norman). – Magdalino, Maritime Neighborhoods 211-212. – Magda-
lino, Grain Supply 43-46. – Müller, Getreide. – Kislinger, Pane. – Durliat, L’ap-
provisionnement 26-27. – Prigent, Rôle des provinces d´Occident. – Kislinger,
Lebensmittel. – Wade, Maritime Cults 269. – Several finds of shipwrecks at
Yenikapı / Harbour of Theodosios (e. g., YK 1, 5, 11, 14, 23, 24; see Pulak / In-
gram / Jones, Yenikapı 105-110 and n. 13. – Jones, Yenikapı 12. – Jones, Cargo
Vessel. – Kocabaş, Shipwrecks 109-111 distinguishes between »sea going trad-
ers« [YK 3, 8, 15, 17-22] and »local trading vessels« [YK 6-10, 12, 14, 31-32]),
proving its commercial significance (and those of most other harbours of the
city) for their subsistence. On the basis of the »local trading vessels« the im-
portance of the local subsistence in the Mediterranean is proven, as stated by
Horden / Purcell, Corrupting Sea 143-152. 365-377. – The wines from all parts
of the empire served at the banquet on the occasion of the crowing of Justin II
were probably brought to Constantinople by ship (Corippus, In Laudem Jus-
tini III 83-93, 96-102, cf. I 109-111; Kislinger, Weinhandel 141-147; about wine
merchants in antiquity s. Brockaert, Navicularii 266-268). – According to Güns-
enin, »City« Harbours 104, from the 7th to the 13th c. local supplies were not
predominantly unloaded at the harbours on the Golden Horn, but mostly at the
harbours on the Propontis coast. – For the later centuries, see Jacoby, Mediter-
ranean Food and Wine, and Günsenin, Ganos, and Jacoby, Mediterranean Food
and Wine; Günsenin, Ganos and Howard-Johnston, Commerce à Byzance 337-
338; about emporoi (traders) and kapeloi (small local merchants) in antiquity s.
Broekaert, Navicularii 257-258. – On archaeological finds of amphorae in the
area of the Theodosios Harbour (probably for wine), see Günsensin, Harbours
and Shipbuilding 420 with fig. 3 (YK 12). – The trade privileges of the branch
of Monemvasia in Pegai (today Karabiga, 56 km west of Bandirma), dating from
1328 and c. 1363-1373, testify to the continuous role of Constantinople as a
maritime trade centre: Kislinger, Zweite Privilegurkunde. – On the significance
of the sea trade in general, see Necipoğlu, Byzantine Economy and the Sea
437-438: »Maritime trade was more profitable than overland trade […] the
sea always remained a major factor fostering trade and a generally flourishing
economy of Byzantium«.
10 Notitia urbis Con­stan­tinopolitanae 233-234, 15-17. English translation in Mat-
thews, Notitia 90. – Mango, Triumphal Way 187-188 (appendix: The Situation
of the Strategion). – Drakoulis, Functional Organization 153-182 merely pro-
vides a wordy presentation of the source and relevant literature. – Kislinger,
Eugenios-Tor 728. – Kislinger, Lebensmittel 311. 314-315. – Berger, Regionen
und Straßen 384-385. – Westbrook, Forum of the Strategion. See also n. 45. –
For comparisons from the Aegean in Antiquity, see Bouras, Geography of Con-
nections 214-215.
11 Kon­stan­tinos Porphyrogennetos, De cer. II 51 (III 394-398 Feissel). – Marcellinus
Comes, Chronicon ad a. 431 (15 Croke) concerning a ceremonial visit by The-
odosios II could also refer to the southern horrea. – Westbrook, Forum of the
Strategion 10.
10 On Better and Worse Sites  |  Ewald Kislinger
somewhat confused history in the Patria Kon­stan­tinoupoleos,
sailors traded grain in the square; a bronze bushel that was
used (earlier) as a standard weight was placed on an arch-
way as a warning to always sell at the correct value of gold
coin (nomisma) 17. The Empress Eirene had halls (triklinous)
built at the bakery (or bakeries) of the Lamia / tēs Lamias tou
pistoreiou (or ta pistoreia) – again a local reference to grain
is made – which, according to the source context, served as
public outlets 18.
In addition to charitable facilities, other establishments
flourished near the harbours, as in different times and differ-
ent places: establishments of commercial hospitality, public
houses with their range of alcoholic drinks for sailors and
dockworkers, sometimes also associated with prostitution 19.
At the same time, state authority had a presence at the
harbours: officials of the eparchy exercised control, such as
the limenarchoi (harbour masters), and levies such as the
limenatikon and skaliatikon (from skala, landing stage) were
collected. The parathalassitai administered justice in disputes
among sailors, official surveys of ships took place at the
Neorion, the l­ogothetes tou dromou was responsible for
privileged foreigners 20.
12 Notitia urbis Con­stan­tinopolitanae 237, 6.  – Berger, Regionen und Straßen
369 does not want to rule out that the Alexandrina granary could have been
situated at the Harbour of Julian because of the boundaries of region IX.
13 Synaxarium Ecclesiae Con­stan­tinopolitanae 78 (Delehaye).
14 Haldon, Comes horreorum. – Durliat, L´approvisionnement 22. 29-30. – Mun-
dell-Mango, Commercial Map 200-201.
15 Berger, Langa Bostani, however, sees a broad harbour function only given until
the beginning of the 8th c., because it has been documented. – Cf. Berger, Re-
gionen und Straßen 373: »Der Theodosiushafen schließlich […] in einer tiefen
Bucht, die im Lauf der folgenden Jahrhunderte verlandete« (The harbour of
Theodosius […] in a deep bay, which silted up in the course of the following
centuries) (scil. after 425), which would have increasingly affected its operability
as a harbour. – Mango, Développement 55.
16 Magdalino, Con­stan­tinople 2 (map). – Janin, Con­stan­tinople 68-69.
17 Patria Kon­stan­tinupoleos II 51 (179 Preger). – Parastaseis syntomoi chronikai
§ 12 (72-74 Cameron / Herrin). On the original location perhaps at one of the
Propontis granaries, see Magdalino, Con­stan­tinople 24 n. 50. – Kislinger, Leb-
ensmittel 308-309 n. 31.
18 Patria Kon­stan­tinupoleos III 85 (246 Preger). In the case of at least this facility
the late antique concept of the pistrina publica is likely – their number was
above average in regions V and IX (i. e., near granaries): Notitia urbis Con­stan­
tinopolitanae 234, 24 and 237, 15. Berger, Regionen und Straßen 384-385. –
Adapted to feeding the poor during the Middle Byzantine period (cf. Volk, Ge-
sundheitswesen 87. 96-97. 130. 182. 208). Magdalino, Con­stan­tinople 25 n. 54
rightly refers to the nearby Myrelaion structure (Mango, Développement 59)
that was in the early 10th c. re-dedicated by Romanos I Lekapenos as a charita-
ble institution, with hospital (xenon), nursing home (gerokomeion) and a daily
distribution of bread (Theophanes Continuatus, Chronographia 430 [Bekker]).
Kislinger, Hospitals, in: Daim, Brill´s History and Culture of Byzantium 469-471.
19 On a state inn for arriving travellers at the Harbour of Julian, built by Isaac II An-
gelos, see Niketas Choniates, Historia 445, 19-23 (van Dieten), where the xeno­
docheion (on this term, see Kislinger, Kaiser Julian 373-378; Kislinger, Lodging
346-347) is denominated as pandocheion, thus in an archaising way. – General
observations (Antiquity), see Rauh / Dillon / Davina-McClain, Ochlos nautikos;
Stasolla, Strutture per l´accoglienza (on Rome, Palermo, Cagliari, Naples, Pisa,
Gaeta); Byzantium: Kislinger, Lebensmittel 310 n. 47 and Kislinger, Reisen 372-
374 n. 188. 194; Veikou / Nilson, Ports and harbours 268-269; entirely unschol-
arly: Stavroulaki, Seaports 28-30 (inns, public houses at / near harbours). On
taverns in the area of Heptaskalon, see Preiser-Kapeller, Heptaskalon, in this
volume, Magdalino, Review 261, and earlier Kislinger, Lebensmittel 317 with
n. 97 and 98. – Macrides, Travel, unfortunately, contains no contribution spe-
cifically on the hospitality industry and lodging in the Byzantine sphere.
20 Book of the Eparch / Eparchenbuch 17.3-4 (128 Koder); Peira 218 (Zachariä von
Lingenthal) – Ahrweiler, Fonctionnaires. – Penna, Imperial Acts 21-24. – ODB III
1586-1587. – Makris, Studien 246-247. 251-252.
 In summary, these harbours fulfilled a threefold purpose:
first, they served as shipping centres for handling and storing
goods; second, the harbours and the neighbouring areas
provided a place of rest and recreation for sailors; third, they
were part of a pan-Mediterranean system of long-distance
and regional shipping routes 21. The harbours of Constantino-
ple as a whole were a hub of international dimensions and
an essential element for the importance of the city, and the
Golden Horn was of great significance for this 22. »The sea,
its mariners, and who serviced the maritime trading industry
influenced life in the eastern capital […] Constantinople was
shaped by its maritime setting« 23.
From the Harbour of Julian, a broad street led to the Tetra­
pylon / Anemodoulion, a gate construction that arched over
the intersection with the Mese 24. All around to the south was
a market quarter called artopoleia, where not only bread was
sold, but also snack bars were installed (as they are today in
such an environment), selling fish, cheese, pulses and wine.
Andrew of Constantinople (also Andrew the Fool) chose to
stay here in the tenth century 25. What is relevant for us, is
that, once again, the surrounding neighbourhood attests
to the continuing functioning of a granary (cf. the Church
of St ­Thecla en tois krithopoleiois, n. 12) at this harbour in
the Middle Byzantine period, thus not only that of the La-
mia-horreum.
The three granaries at the Harbour of Prosphorion on
the Golden Horn, on the other hand, disappeared from the
sources after the sixth century. If the horrea can be equated
with parathalassia apothekai (magazines near the sea), then
the last document from 561 attests to their destruction dur-
21 On Constantinople as a starting point for Mediterranean sea routes, see Kislinger,
Sea Routes 320-322 and Kislinger, Markets and Fairs, Trade Routes 390-393,
both in Daim, Brill´s History and Culture of Byzantium. – On connectivity, see
Kolditz, Horizonte maritimer Konnektivität; Bouras, Geography of Connections;
Avramea, Land and Sea Communications; McCormick, Origins 502-508. 531-547.
593 (map 20.2 shipping routes: simplified segments); Külzer, Pilgerwege und Kul-
torte 183-187. – Heher / Preiser-Kapeller / Simeonov, Vom Lokalen zum Globalen
201-209, on the Orbis Stanford Geospatial Network Model of the Roman World,
which classifies and analyses junctions according to degree-centrality (intensity of
connections), betweenness-centrality (importance within the whole system) and
closeness-centrality (distance to other junctions). In the general network, several
clusters exist. Its junctions share a higher degree of connectivity, e. g., the Propon-
tis region and the Black Sea area. Constantinople is the hub of both clusters. – The
compilation of common law orders in seafaring in the Nomos Rhodion Nautikos
(8th c.?) emphasises their empire-wide importance with a need for harbours (Let-
sios, Seegesetz der Rhodier; Howard-Johnston, Commerce à Byzance 313-316.
340-341). – The internationality of Constantinopolitan harbours is demonstrated
by, e. g., small finds from the Theodosian Harbour, such as a gold ring belonging
to Undila (possibly of Gothic origin, 6th c.) and a (Middle Byzantine?) roof tile,
which names amongst others a certain Karelos / Karilos, a name commonly found
in the Latin-speaking provinces of the Western Merovingian sphere: Tsivikis, Epig-
raphy 121-122 and 124-125. – Feuser, Hafenstädte 1-2. 4-5. 15.
22 See Kislinger, Golden Horn, in this volume.
23 Wade, Eternal Spirit of Thalassa 54. – Criticism of sailors’ negative influence on
urban society in Libanius, Oratio XI 38 (I/2 448-449 Foerster); Sokrates, Historia
Ecclesiastica VI 15, 11 (337 Hansen). – Wade, Lock Up your Valuables 53-54. 71.
73. 75. – Cheynet, Poids politique des marins. – Kolditz, Horizonte maritimer
Konnektivität 73 n. 75. – For traders in Antiquity cf. Feuser, Hafenstädte 284-286.
24 Patria Kon­stan­tinupoleos II 46/46a (174-175 Preger). – Cf. Parastaseis syntomoi
chronikai § 40 (106-110 Cameron / Herrin). – Berger, Untersuchungen 315. –
Anderson, Anemodoulion. – Kislinger, Lebensmittel 313-314. – Bauer, Stadt,
Platz und Denkmal 351-353. – Bread was sold in the Artopoleia area, not grain,
so Howard-Johnston, Commerce à Byzance 329 (with insufficient knowledge of
relevant literature). The Middle Byzantine grain trade took place at tou Amastri-
ing a fight of the circus factions, the Greens and the Blues 26.
This would match the information that the markets for mar-
itime merchandise (agorai tōn thalassiōn emporeumatōn) at
Neorion were relocated to the Harbour of Julian by Justinian I
(reg. 527-565) 27. The Harbour of Neorion continued to exist,
and Emperor Leontius even had it dredged in 698 28. During
the months of fighting between Emperor Anastasius II and
the usurper Theodosius in 715, Neorion functioned as the
base for the imperial fleet, alongside Hagios Mamas on the
Bosphorus 29. In general, it can be assumed that the harbours
on the easily defensible Golden Horn 30 served the navy since
the first Arab siege of 668/669 31, including a shipyard or
shipyards. Ensuring the secrecy of armaments and weapons,
especially that of Greek or liquid fire (hygron pyr) 32, was cer-
tainly the primary motive in shielding this military complex
from spying eyes, which was made easier by the concentra-
tion of civilian seafaring elsewhere. Berger is incorrect in be-
lieving that the Harbour of Julian / Sophia (later Kontoskalion)
on the Sea of Marmara had been a »naval base« in the Mid-
dle Byzantine period 33. This only applies to the Palaeologan
period (see below and n. 133).
From the sixth century, the harbours on the Propontis
were preferred for handling goods for the simple reason that
they were closer to the commercial and political centre, which
extended parallel to the south coast of the city along an east-
west axis, with the Mese as its main artery 34. Several squares
opened along this portico-lined street, such as the Forum
of Constantine or the Forum Tauri 35, to mention only the
most important. This is where many traders and craftsmen
had their shops 36, and around the Forum Tauri in region VIII
anou (see above n. 16), near the Propontis horrea. – In Berger, Regionen und
Straßen 397 (fig. 5). 405-406, the hypothetical road from the Harbour of Ju-
lian (Kumkapi) contradicts an intersection with the Mese (and continuation by
the Makros Embolos) at the site of the Tetrapylon (on this denomination, see
Berger, Toponyms of Constantinople 164).
25 Vita Andreae Sali 28, 38, 92-94 (Rydén). – Cf. Parastaseis syntomoi chronikai
§ 13 (76 Cameron / Herrin): Artotyrianon. – Mango, Développement 55. – Mag-
dalino, Con­stan­tinople 22-23.
26 Ioannes Malalas, Chronographia XVIII, 135 (424 Thurn). The mansion of An-
dreas en tō Neoriō was also destroyed (loc. cit.).
27 Parastaseis syntomoi chronikai § 72 (152 Cameron / Herrin). – Patria Kon­stan­
tinupoleos II 68 (188 Preger).
28 Theophanes, Chronographia 370 (de Boor).
29 Theophanes, Chronographia 385-386 (de Boor). – Ahrweiler, Byzance et la mer
430-431.
30 For the blockade chain attached to this entrance as needed, see Pryor / Wilson,
Chain. – Guilland, Chaine (= Guilland, Études de topographie II 121-146). –
Kedar, Chains 5-6. 22-24. 26. – Makris, Studien 182-184. – Kislinger, Golden
Horn, in this volume.
31 Jankowiak, First Arab Siege.
32 Haldon, »Greek Fire« Revisited.
33 Berger, Häfen 83. See the earlier reference in Berger, Häfen von Byzanz und
Konstantinopel 114: »… der Hafen der Sophia, in dem Zeit seines Bestehens
immer die byzantinische Kriegsflotte lag«. Cf. now Heher, Harbour of Julian, in
this volume, and Günsenin, Harbours and Shipbuilding 417.
34 Another example for the interaction between harbours and urban life is given
by ancient and Byzantine Syracuse (Castagnino Berlinghieri, Portualità di Sira-
cusa), where its centre, Ortygia, also from its peninsular shape is comparable
to Constantinople.
35 Bauer, Stadt, Platz und Denkmal 148-268. – Müller-Wiener, Bildlexikon 255-
265. – Berger, Tauros e Sigma. – Barsanti, Il foro di Teodosio.
36 Detailed information goes beyond the current subject, see Mundell-Mango,
Commercial Map 199-203. – Thomov / Ilieva, Shape of the Market. – Schreiner,
Costantinopoli 108-111.
On Better and Worse Sites  |  Ewald Kislinger 11
(and VII?) were two macella 37. To the west, the Mese and its
extensions were connected to the long-distance overland
routes through the Balkans (which supplied merchandise
from the surrounding area) 38 and also allowed the connection
of the Harbour of Theodosius to the urban infrastructure. In
its eastern part, the Mese led to the »government quarter«
with the Imperial Palace, Hagia Sophia and Hippodrome, all in
close proximity and complementary in function. This is where
secular and religious power manifested itself physically and
where it was staged ceremonially before, and also sometimes
with, the public 39. This quarter connected  – slightly out of
alignment to the south-west – to the Acropolis, with its tem-
ples, two theatres and the lusorium of ancient Byzantion 40,
hence represented continuity adapted to the new dimensions,
rather than change.
This southward shift of the centre from the Golden Horn
to the Sea of Marmara resulted from the interaction with the
function of the Mese as an economic focus and attraction
(fig. 1). It is unlikely that the process was planned as compre-
hensively as the results show. The authorities wanted only to
accomplish the structural growth of the city initially enforced
in the fourth to fifth centuries, using the newly developed
areas, not least for the new harbours of Propontis. Harbours
and granaries are essential utilitarian facilities in the larger
concept of large-scale city planning, beginning with the The-
odosian dynasty 41. The old residential areas on the slope of
the Golden Horn (regions  V-VII und X) 42, including utilities
(such as market halls / macella 43 and water supply 44), were
supposed to be preserved. Accordingly, as mentioned at the
beginning, a bipolarity existed (examples are the dispersion
of macella) from the fourth to the sixth century, but it is ques-
tionable whether this was completely stable.
Conveniently located, not overcrowded quarters of a large
city with good infrastructure, near the political and cultural
37 Notitia urbis Con­stan­tinopolitanae 236, 17. – Mundell-Mango, Commercial
Map 193-194. – Kislinger, Lebensmittel 312-313. Leomakellon and Dimakellon
(also mentioned in Kleinchronik 14, n. 1a, [Kleinchroniken 130 Schreiner]) and
tou Makellou (at the Forum of Constantine: Sokrates, Historia Ecclesiastica I
38, 9 [89-90 Hansen]) and ta Makellou (the existence [?] of which is based only
on Manuscript D of the Vita Andreae Sali. ch. 2 [18 Rydén, app. crit.] and with
reference thereon Janin, Siège de Con­stan­tinople 29) must – contrary to Berger,
Untersuchungen 184 and 515 – be distinguished.
38 Kislinger, Verkehrswege und Versorgung (with further literature). – Külzer, Ost-
thrakien 192-202.
39 For a short selection, I refer to Dagron, Déroulement des courses.  – Hippo-
drom / Atmeydanı. – Bauer, Visualisierungen von Herrschaft. – Featherstone, Der
Große Palast. – McCormick, Eternal Victory. – Majeska, Emperor in his Church. –
On out-reaching productions, see Berger, Straßen und Plätze.
40 Berger, Regionen und Straßen 357-360.
41 Magdalino, Renaissances 58-59.
42 Notitia urbis Con­stan­tinopolitanae 233-236. 237-238. English translation in
Matthews, Notitia 89-91. 93-94. – Berger, Regionen und Straßen 377. 382-
383.
43 Notitia urbis Con­stan­tinopolitanae 234. – Kislinger, Lebensmittel 312-314. –
Mundell Mango, Commercial Map 193-194. – Berger, Regionen und Straßen
385-386. – On the concept generally and its evolution, see De Ruyt, Macellum. –
Cf. Lavan, Retail and Regulation 342-343. 346. 367 on such facilities elsewhere.
44 Hadrian’s water system supplied ancient Byzantion, that of Valens also encom-
passed the area around and south of the Mese and (by means of the Cistern
of St Mocius) in the southwest of the city: Bono / Crow / Bayliss, Water Supply. –
Crow / Bardill / Bayliss, Water Supply, esp. 9-20. – Crow, Infrastructure 268-279
12 On Better and Worse Sites  |  Ewald Kislinger
hot spots were always attractive in history for the rich and
powerful, who had the necessary means and prerequisites to
live there. In the area south of the Mese where the ground
slopes down to the Sea of Marmara, a specific Byzantine
attraction was a wide and beautiful view 45. Corippus praised
the palace of Julian II and Sophia above the Harbour of Ju-
lian, which was thus removed from its noise and odours:
»Welcome to the noble couple was the place where they
used to observe the surging of the sea and the curved ships
who brought all goods from the continents« 46. Justin and
his consort were not among the first at this site, who had
moved to this area and further increased its value; they also
found followers 47.
In addition to the basic economic and sometimes military
function of harbours, in Constantinople as elsewhere, it is
also important to consider their representative ceremonial
role. This ranged from imperial landings and the reception of
state guests to the transfer of relics 48. Thanks to the contri-
butions by Simeonov and Heher in this volume and elsewhere,
a detailed analysis can be omitted here. A change can be
observed in this area as well: the focus was relocated from
the ancient Prosphorion / Strategion area to Hebdomon in
Early Byzantine times, then to the Boukoleon Harbour, and
in the Late Byzantine period, also the Golden Horn became a
preferred locus for all the above mentioned representational
and trade-related tasks 49.
Around 540, Byzantine Constantinople reached its highest
population of just under half a million people 50. Analogous
to the expansion of the empire (renovatio imperii romani,
with expansions into the central Mediterraneum) a quick and
long-lasting steep decline happened, not only in a territorial,
but also a demographic sense. This had been caused primarily
by constant war on all fronts (against Sasanids and then Ar-
abs in the east and southeast, Avars and Slavs in the Balkans,
fig. 1. – Crow, Ruling the Waters. – Sürmelihindi et al., Byzantine Water Man-
agment. – Mango, Water Supply. – Berger, Regionen und Straßen 379-381.
45 Saliou, Traité d’urbanisme § 52-56 (72-75). – Saliou, Lois des bâtiments 238-
246. – Velenis, Wohnviertel 229. – Dark, Eastern Harbours 157. – Grünbart,
Inszenierung 74. 90-92.
46 Flavius Cresconius Corippus, In laudem Iustini I 109-111. – Cf. I 101-103 (39
Cameron). In general Libanius, Oratio XI 37 (I/2 448 Foerster). On the compa-
rable later expansion of the imperial palace complex southward and the emer-
gence of the Boukoleon part including harbour, see Heher, Boukoleonhafen
123-124.
47 Nikephoros Phokas the Elder and his son Bardas followed after Justin II in the
position above the Harbour of Sophia (Leon Diakonos, Historia V 5 [83-84
Hase]. – Niketas Choniates, Historia 445). The Empress Eirene resided in the
Palace of Eleutherios (Vita Basilii Minoris III 36, 1), near the one of Arcadius
(Notitia urbis Con­stan­tinopolitanae 237, 7). – Mango, Développement 59. –
Magdalino, Maritime Neighborhoods 216.
48 Heher / Simeonov, Ceremonies by the Sea 223-227. – On the image of the har-
bour in literary comparison, see Chrysos, Limen. – In general: Bauer, Stadtver-
kehr in Konstantinopel; Berger, Straßen und Plätze. For parallels from antiquity
Feuser, Hafenstädte 271-274.
49 Heher / Simeonov, Ceremonies by the Sea 227. 230-233. 235-236. – Vučetić,
Repräsentative Aspekte von Häfen 135-140.  – Schreiner, Brautgedicht and
Pseudo-Kodinos, Traité, chap. 12 (286-287 Verpeaux).
50 Koder, Lebensraum 117-118. – Low estimate of 375 000 inhabitants before 541
given by Schreiner, Costantinopoli 81-83; higher estimate of 600 000 given by
Durliat, Ville antique 232-275 n. 210. – Jacoby, Population.
Fig. 1  The harbours and economic
axes of Constantinople and their loca-
tions over time. – (Map E. Kislinger /
J. Preiser-Kapeller).
Lombards in Italy) 51 and widespread pestilence 52. When the
Justinian plague ebbed away after a massive eruption in 743-
750 53, which affected Constantinople in 747-748, the city
probably reached its population low point. Although 40 000
inhabitants may be too pessimistic an estimate 54, even if dou-
ble that is estimated, the city would have lost more than 80
per cent of its population level of 540. The fallout of this for
the cityscape was, on the one hand, the contraction of set-
tlement, especially on the Mese axis (the better position was
still preferred) and the transverse from the Harbour of Julian
via Makros Embolos (»great shopping street«; now Uzunçarşı
Caddesi) to Perama 55. On the other hand, sparsely populated
and deserted areas created a spatial surplus that also had its
advantages. As the harbours were no longer able to operate
at full commercial capacity due to lower demand, their num-
51 Overviews offer pars pro toto: Whitby, Maurice. – Pohl, Avars. – Zanini, Italie
bizantine. – Kaegi, Early Islamic Conquests. – Stratos, Byzantium. – Eickhoff,
Seekrieg und Seepolitik 9-50.
52 Stathakopoulos, Famine and Pestilence. – Meier, Pest, and the chapters in Mei-
er’s volume by W. Brandes (201-224) and K.-H. Leven (11-32). – Congourdeau,
Pandémies. – Meier, »Justinianic Plague« rightly criticises a recent attempt to
minimise the effects of this pandemic, but overemphasises the importance of
Meier, Pest. – On Con­stan­tinople, see Kislinger, Pane 279-293.
53 Stathakopoulos, Famine and Pestilence 379-386.
54 Mango, Développement 53-54.
55 Mundell Mango, Commercial Map 197 fig. 31.
56 Notitia urbis Con­stan­tinopolitanae 233, 11-12 speaks of Strategium, in quo
est forum Theodosiacum et obeliscus Thebaeus quadrus. English translation
in Matthews, Notitia 90. – Parastaseis syntomoi chronikai, § 24 (84-86 Cam-
eron / Herrin) differentiate a large and small Strategion. Mango, Triumphal Way
187: »It can be provisionally suggested, that we have here a civic forum flanked
ber could be reduced for this purpose, or, as already stated,
diversification of use became possible: the Neorion Harbour
on the Golden Horn thus passed to the navy for centuries to
come (see above).
In the hinterland of the neighbouring Prosphorion Har-
bour to the east, the Strategion – still one of the city’s great
squares in the fifth century 56 and a centre for the regions
of the lower Golden Horn – retained its function as a cattle
market for the time being 57. It was only under Constantine V
(reg. 741-775) that this was transferred to the Forum Tauri 58,
that is, at the time of the city’s population low point. The
decisive factor was probably the question of local supply 59 of
the densely populated zone in the area of the Mese and to
its south, which Constantine V focused on regarding the city
planning 60, which overrode the hygiene aspect.
by a smaller marketplace«. – Mundell Mango, Commercial Map 192. – West-
brook, Forum of the Strategion 5-7. – Bauer, Stadt, Platz und Denkmal 224-
228. – An arch or entranceway crowned with a Fortuna (Marcellinus comes,
ad annum 510 [35 Croke]. Mango, Développement 19 n. 32) was considered
by the Patria Kon­stan­tinupoleos I 51 (141 Preger) to be the Arch of Urbicius
(opening to his nearby house, see n. 81) on the speculative Byzas Wall.
57 For a similar use of the lower Agora near the harbour in Ephesos, see Foss,
Ephesus 63. 82 fig. 12.
58 Patria Kon­stan­tinupoleos III 149 (263-264 Preger). – Magdalino, Renaissances
75; cf. even earlier Kislinger, Lebensmittel 314-315 and Kislinger, Von schlech-
teren und besseren Lagen, in: Daim, Häfen 12.
59 Nikephoros Patriarches, Breviarium ch. 85 (160 Mango) on the rich supply of
markets under Constantine V.
60 Magdalino, Con­stan­tine V, 10-11. – Cf. (2001). – Magdalino, Maritime Neigh-
borhoods 213 n. 28.
On Better and Worse Sites  |  Ewald Kislinger 13
 Since the water supply via the Valens Aqueduct to the
nymphaeum maius was disrupted in the course of the Avar
siege of 626 61, the reservoirs 62 that this had fed could only
be filled by rain water, but this was apparently sufficient given
the reduced population. Only as a result of the state-ordered
resettlements from various provinces in 755 63 did the situa-
tion threaten to become precarious in the event of prolonged
drought. When that occurred in 766, Constantine  V had
the Valens Aqueduct repaired, for which he now had to
bring builders to Constantinople from the Pontus, Hellas, the
islands and from Thrace 64. In distant Naples, the measure
became the fairytale-like expulsion of a dragon from the
aqueduct, which had previously carried off many inhabitants
with its exhalations (see n. 81 on the miasma concept). Lack
of water due to blocked or interrupted supply lines, resulting
in a lack of hygiene, encouraged the spread of disease, espe-
cially the plague outbreak of 747-748, and the intervention
of the emperor was according to this legend brought nearer
in time and combined with this 65.
Trade in pigs at the Forum Tauri 66, to which the animals
were driven up from the Harbour of Julian 67, and that with
horses (brought from Thrace?) at the Amastrianum 68 are
further indications of mercantile concentration in appreciated
and promoted residential areas. In contrast, the ambience
around the palace was supposed to be dominated by the
fragrances of musk, frankincense and myrrh, products of
the druggists from the Milion to the Chalke Gate 69. An Early
Byzantine slaughterhouse at the Forum of Constantine was
probably relocated due to the unpleasant smells 70.
Less noble districts, which had become more remote
due to the negative demographical growth of Constan-
61 Nikephoros Patriarches, Breviarium ch. 85 (160 Mango).
62 Crow / Bardill / Bayliss, Water Supply 20.
63 Theophanes, Chronographia 429 (de Boor). The measure is certainly to be seen
as a reaction to the loss of population caused by the plague outbreak of 747-
748.
64 Theophanes, Chronographia 440 (de Boor). – Magdalino, Water 132. – Mag-
dalino, Renaissances 72-73. 75. – Perhaps the expulsion of various monastic
communities from their monasteries and their re-dedication as barracks for new
elite units (Theophanes, Chronographia 437 [de Boor]). – Magdalino, Con­stan­
tine V, 3. 6. 12) was not solely ideologically motivated, but partly due to lack
of accommodation and construction workers, or perhaps due to earthquake
damage (Magdalino, Renaissances 74). – The reproach of the Emperor for sell-
ing liturgical objects in order to finance the construction costs of houses, baths
and theatres also points in this direction (proceedings of the Second Council
of Nicaea 787: Mansi, Collectio XIII 333 A-B, see ACO series secunda, volume
tertium, pars tertia 756, 9-11 (Lamberz / Dubielzig).
65 Gesta episcoporum Neapolitanorum 422-423 (Waitz). – Acconcia Longo, Agio-
grafia e narrativa tra Oriente e Occidente 245-248. – On this plague wave, see
Stathakopoulos, Famine and Pestilence 384-385.
66 Book of the Eparch / Eparchenbuch 16.2 and 16.3 (124-126 Koder). – Book of
the Eparch / Eparchenbuch 15.5 (124 Koder) also testifies to trade in lambs from
Easter to Pentecost.
67 Patria Kon­stan­tinupoleos II 46a (175 Preger). – Kislinger, Lebensmittel 313-314.
68 Book of the Eparch / Eparchenbuch 21.3 and 21.8 (136-138 Koder).
69 Book of the Eparch / Eparchenbuch 10.1 (110 Koder).
70 Sokrates, Historia Ecclesiastica I 38. 8-9 (89-90 Hansen). – Kislinger, Lebensmit-
tel 314.
71 On this equalisation earlier Schneider, Mauern und Tore 77. – Asutay-Effen-
berger, Porta veteris rectoris 133.
72 Kislinger, Lebensmittel 316.  – Asutay-Effenberger / Effenberger, Eski Imaret
Camii 23-24. – Berger, Ufergegend 153.
14 On Better and Worse Sites  |  Ewald Kislinger
tinople, such as Leomakellon near the coast at Basilike pyle
(Unkapanıkapısı) 71 on the Golden Horn 72 (or rather more
precisely at the Heptaskalon 73) or around the site of the
Strategion, were better suited for such purposes or also
trades with an associated fire risk, such as glass-blowing:
»However, if any necessity required it within the cities, the
hyalourgoi must operate in isolated locations away from the
residential areas« 74. Such a workshop (ergasterion hyelopses­
tikon) on the steep street (Dikymbalos) to the Hagia Sophia
obviously complied with the legal requirements; nevertheless,
a fire broke out from it that raged to the Chalkoprateia
Church 75. It was primarily for religious reasons that the Jew-
ish community established a separate settlement. However,
in 1044 (?), or at least before 1082, they had to move from
the south bank of the Golden Horn to the northern one 76.
The evil odour emanating from the tanneries no doubt also
motivated the change 77. This, too, fits for the time before
the image of a predominantly commercial zone in the hin-
terland of the devalued and re-designated harbours on the
Golden Horn.
As a reason why it ever came to this image, the plague
was made responsible 78. The thousands of dead in the first
wave in 542 were taken to the Golden Horn, stacked on
the other bank at Sykai, buried in mass graves 79 or thrown
into the sea 80. In other epidemics similar procedures may
have been used. The associated miasma of the area 81 is
implicitly linked to a renewed outbreak of the plague in 698
when the Neorion Harbour was dredged 82. The causality can
also be modified. In the poor, overpopulated tenements of
adjoining urban areas, the plague  – probably reaching the
city by sea, with harbours being the typical gateway of a
73 Vita der Theodosiae 131 (Gedeon). – On the Heptaskalon, see Preiser-Kapeller,
Haptaskalon, in this volume.
74 Julianus Ascalonites § 11.1 (Saliou, traité d’urbanisme 40-41). – Hexabiblos II 4,
19 (117-118 Pitsakis). – Velenis, Wohnviertel 227.
75 Invention des reliques et miracles de Ste Photine (BHG 1541 m), ch. 9 (122-123
Halkin). – See Talbot, Photeine 101 n. 52. – Henderson / Mundell Mango, Glass
344-346. – Mundell Mango, Commercial Map 202-203, n. 119. – Mango, Tri-
umphal Way 188, still locates some Ottoman workshops for glass production
slightly northeast.
76 Jacoby, Quartiers juifs 170-171. 181-183. – Jacoby, Jews 223-225.
77 Benjamin de Tudela, Itinerarium 24 (Adler).
78 Magdalino, Maritime Neighborhoods 217-219.
79 Prokopios, Bella II 23, 9-11 (I 257 Haury / Wirth).
80 Ioannes Ephesius, Vitae sanctorum Orientalium 89 (Brooks).
81 Gen. Pseudo-Athanasius Alexandrinus, Quaestiones ad Antiochum, erot. 103
(PG 28, 661 A-B). – Anastasios Sinaites, Questiones et responsiones, erot. 66
(118-119 Richard / Munitiz). – Aetius Amidenus, Libri medicinales V 95 (II 80-81
Olivieri). – Paulos Aiginetes, Epitome iatrike II 34 (I 107-108 Heiberg). – The rel-
evant sources quoted by Magdalino, Maritime Neighborhoods 218, n. 65 and
66 are in parts from outdated editions. – For further examples of the miasma
idea, see Acconcia Longo, Agiografia e narrativa tra Oriente e Occidente 247
n. 61.
82 Theophanes, Chronographia 370 (de Boor).  – Berger, Häfen 80-81 follows
Theo­phanes’ arguments and sees the cause of the plague wave of 698 in silt
and waste, without a mention of Stathatkopoulos, Famine and Pestilence,
esp. 364-365. Considering the knowledge of transmission paths of the plague,
the common opinion that waste was the real catalyst of the pandemic seems
unlikely. However, this was rather based on the negative image of the urban
quarter since 542. See on this Magdalino, Constantinople 99; Magdalino, Mar-
itime Neighborhoods 218-219 and Kislinger, Von schlechteren und besseren
Lagen 12-13 in Daim, Häfen (German version [2016] of the present article).
pandemic 83  – will have raged more fiercely than in richer
neighbourhoods with higher sanitary standards 84. It was log-
ical in a catastrophic situation to bury the dead nearby. The
interests of the local survivors were of no great concern, their
demographic and public weight shrank due to the epidemics,
the already second-rate area now became the slum of the
city: »The Golden Horn took a long time to shake off its bad
reputation …« 85.
The first signs of change again became evident in the
tenth century 86. In the Book of the Eparch, the makelarioi are
instructed to buy (and to slaughter) sheep until the beginning
of the pre-Easter Lent at the Strategion  – from which the
bronze sculptural decoration had been taken away a few
decades earlier 87. Only the trade with lambs remained at the
Forum Tauri from Easter to Pentecost 88. In a high-turnover
period, the market was left close to the customer, whereas
otherwise it was removed from the centre again. The measure
was taken for hygienic reasons, and is likely to be connected
with the noticeable increase in population as the Empire be-
gan to tackle the last wave of Arab attacks at sea and soon
asserted itself against the Bulgarians 89. It is significant from
the supply logistics point of view that in 960, when prepa-
rations for the then successful landing on Crete were made,
additional grain needed for this purpose was to be bought
in from the east and west 90. About eighty years later, already
the mere supply for the metropolitan population, which had
to be secured in the face of a shortage, necessitated a similar
course of action 91.
Demographic growth was also manifested in urban de-
velopment, and areas that were lying fallow for much of
the time gained in interest. After Urbicius, author of a tak­
tikon under Anastasios (reg. 491-518), whose house lay in
the Strategion (after 548 it became a Syrian monastery 92),
we encounter with Antonios, a prominent resident near the
Neorion wharf 93. Antonios was Patrikios in the time of Mi-
chael III (reg. 842-867) and owned an elegant house in the
old harbour district, the private bath of which he opened
for charitable purposes. The group of believers (presumably
a brotherhood), which continued this work, however, was
83 Stathakopoulos, Famine and Pestilence 31. 137-138. – Kislinger / Stathakopou-
los, Pest und Perserkriege 85-93. – McCormick, Bateaux de vie, bateaux de
mort. – Bergdolt, Der Schwarze Tod 35-41.
84 Conrad, Pest. – Dark, Houses 87-89. – Westbrook, Forum of the Strategion 24.
85 Magdalino, Con­stan­tinople 99.
86 Book of the Eparch / Eparchenbuch 15.1 and 15.5 (122-124 Koder).
87 Patria Kon­stan­tinupoleos II 61 and III 24 (184. 221 Preger). – Bauer, Stadt, Platz
und Denkmal 227-228. – Bassett, Urban Image 242-244. – Magdalino, Water
137-138.
88 Book of the Eparch / Eparchenbuch 15.1 und 15.5 (122. 124 Koder). – Mundell
Mango, Commercial Map 199-200.
89 Tougher, Leo VI 164-193. – Eickhoff, Seekrieg und Seepolitik 258-261. – Ste-
phenson, Balkan Frontier 18-23. – Kislinger, Verkehrsrouten 164-165.
90 Theophanes Continuatus, Chronographia 479 (Bekker).
91 Ioannes Skylitzes, Synopsis 400 (Thurn).
92 Patria Kon­stan­tinupoleos III 22 (220 Preger). – Ioannes Ephesius, Vitae sancto-
rum Orientalium 683 (Brooks). – PLRE II 1190. – Janin, Con­stan­tinople 400. –
Berger, Untersuchungen 404-405.
93 Synaxarium Ecclesiae Con­stan­tinopolitanae 935-936 (Delehaye).  – PmbZ I
no. 558.
financially weak, thus corresponding to the social image of
the district, the diaconia tēs Theotokou en tō Neoriō was
in decline 94. An imperial prospect was found in the tenth
century with Romanos Lekapenos. However, it was originally
intended to demolish the existing buildings to make way for
the emperor’s palace. A vision of St Mary ordered a halt to the
work; the bath was renewed and, as a metochion, attached
to a monastery 95. As a former commander of the navy 96, it
was not accidental that Romanos wanted to settle near the
Neorion and evidence suggests that he succeeded 97: a palace
situated on a terrace above the Golden Horn could be that
of Romanos. It was later named after the families of Botanei-
ates and Kalamanos, and passed to the Genoese in 1192 98.
Regarding the development of ownership and the locality, it
would then be a parallel to the residence of Justin II at the
Harbour of Julian / Sophia.
The Monastery of Manuel was financially supported by
Romanos I. It had three skalai below »his« palace 99 and was
not the only monastery that had possessions along the banks
of the lower Golden Horn. Almost all of them, as Magdalino
could prove 100, emerged in the tenth and eleventh centuries,
and received their endowment and validation in this time.
There was still sufficient space available in the expanding
Constantinople, not least in the area between the Phos-
phorion Harbour and the ferry to Perama. For a long time a
problem area of the city, this stretch of shore and the inland
areas were revitalised. Michael VI (reg. 1056-1057) even went
to renew the overgrown Strategion, a logical step »in a part
of the city that was returning to importance«, which earned
him ignorant mockery 101. The ambience of the Golden Horn
did not count as one of the city’s best areas.
Exceptions such as the palace of Despot Constantine An-
gelos or the house of Sebastocrator Isaac Comnenos (later
the Monastery of Christos Euergetes) 102 prove the rule for
the time being. These were also located on the coast north-
west of Perama, in the upper part of the Golden Horn, which
would then experience a sustained appreciation with the be-
ginning of the Comnenian period. The new dynasty raised the
Blachernae quarter as the new seat of imperial power. This
  94 Magdalino, Con­stan­tinople 34. 106. – Generally on this subject see Magdal-
ino, Church, Bath and Diakonia. Repeated in Magdalino, Water 134-135.
  95 Synaxarium Ecclesiae Con­stan­tinopolitanae 937-938 (Delehaye).
  96 PmbZ II no. 26833.
  97 See the convincing arguments by Magdalino, Con­stan­tinople 94.
  98 Cupane, Traumpaläste 411-426.  – Grünbart, Inszenierung 74-75.  – Dark,
Eastern Harbours 57 (terrace at the Cemal Nadir sokak). – Berger, Ufergegend
162 (western slope of the Acropolis).
  99 Theophanes Continuatus, Chronographia 432-433 (Bekker).  – Magdal-
ino, Con­stan­tinople 91-92 n. 208. – Hesitant but ultimately similar Berger,
Ufergegend 162.
100 Magdalino, Con­stan­tinople 92-93.
101 Ioannes Skylitzes, Synopsis 482 (Thurn). – Magdalino, Con­stan­tinople 57-58.
The presence of sieve makers at the end of the 12th c. on the site indicates
that the Strategion had not really risen: Ioannes Nomikopulos, Ekphrasis 296
(Karpozelos).
102 Magdalino, Con­stan­tinople 89-90 n. 198, 80. – Asutay-Effenberger, Kynegion
District, in this volume rejects the common identification of the Monastery of
Christ the Benefactor with Gül Camii.
On Better and Worse Sites  |  Ewald Kislinger 15
focus formation starts in parallel with another change (which,
however, happened separately despite a relative proximity),
which has a primary interest from the perspective of those
interested in the harbours: the emergence of western com-
mercial settlements further southeast of the Golden Horn.
It is assumed that the relevant contracts and political back-
ground are known and do not need to be discussed here 103.
The essential question is why the choice fell on the Golden
Horn and not the economic centre with the Mese and the
Propontis harbours.
»There can be no doubt, that the establishment of the
Italians increased the commercial importance of the Golden
Horn. But would the Italians have asked for concessions in
this area if it had not been fairly important already to their
business interests?« 104. Indeed, positive arguments could be
provided, such as the proximity to branches of other eco-
nomic operators, mainly Arab traders in Perama or at the end
of the Makros Embolos. Besides familiarity of the Venetians
(and those from Amalfi) with the Neorion Harbour and the
wharf area – first at Neorion, then in Sykai – as sailors in im-
perial service 105. Nevertheless, this approach, even the overall
concept, over-estimates in a central element the political
creative force of the participants from the West. It was the
Byzantine state, which, albeit facing hostile pressure at the
beginning of the Comnenian period, issued trade agreements
and concessions. These were formally expressed in the gra-
cious granting of a privilege. Before doing so, concrete inter-
ests of the empire were taken into account or even preceded.
The recipients, first Venice and then Pisa (1112), were at best
able to express their wishes, that is all they could, and there
was no question of free choice on their part 106. The threefold
Genoese proposal of 1155 (settlement west of the Venetians
was preferred, second choice was the Prosphorion district,
otherwise beyond the city in Sykai / Pera 107) shows the still
narrow limits of foreign influence on the choice of location.
103 A brief selection includes Lilie, Handel und Politik. – Pacta veneta 992-1198. –
Pacta veneta 1265-1285.  – Nicol, Byzantium and Venice.  – Banti, Amalfi,
Genova, Pisa e Venezia. – Balard, Romanie génoise. – Origone, Bisanzio e
Genova. – Balard, Amalfi et Byzance. – Skinner, Medieval Amalfi. – Italiens à
Byzance. – Jacoby, Venetian Quarter. – Maltézou, Quartiere veneziano.
104 Magdalino, Maritime Neighborhoods 219.
105 Magdalino, Maritime Neighborhoods 220. – Magdalino, Con­stan­tinople 95.
98-99. – On Sykai, see Müller-Wiener, Häfen 12-13.
106 In contrast to Magdalino, Maritime Neighborhoods 220.
107 Sanguineti / Bertolotto, Documenti 346.  – Magdalino, Maritime Neighbor-
hoods 221-222.
108 Reinert, Muslim presence in Con­stan­tinople. Against Reinert and with Magda-
lino, Con­stan­tinople 98, I am of the opinion that the mosque grew out of an
earlier merchants’ accommodation near the Makros Embolos (mitaton: Book
of the Eparch / Eparchenbuch 5.2 [Koder]). – Pontani, Note 302-304 seeks to
derive an equation between synagogion (recte mosque) and mitaton from
Niketas Choniates, Historia 553, 91-95 (van Dieten). The source merely states
that the entire area was referred to (in demodes dialektos) as mitaton, which
reflects the far longer existence of the merchants’ accommodation compared
to the more recent mosque.  – Cf. Di Branco, Ismailiti a Bisanzio 119-120,
who loc. cit. also proves that Pontani has overlooked another Niketas pas-
sage (Historia 525, 19-20 [van Dieten]), in which synagogion clearly refers to
a mosque. – Turchetto, Mitaton 269-270. 283 follows Pontani de facto, in
wanting to situate the mitaton (correctly) within the sea walls (272), but at the
same time putting it slightly southeast of the Church of St Eirene of Perama
(271 fig. 2) without any evidence for this. – Cf. Jacoby, Venetian Quarter 159.
16 On Better and Worse Sites  |  Ewald Kislinger
An Italian preference for the Golden Horn cannot be deduced
from this, Genoa orientated itself only on the current market
situation in the literal and figurative sense.
Similarly, the Arab proximity to Perama 108, arranged some
70 years before, will have prompted Venice to accept the
granted assignment of land on the Golden Horn – and thus
away from the prosperous Mese and the Propontis harbours.
But Muslim trade partners had been placed there by the
Byzantine state authority, so they were still doing better than
the merchants of the Rhus, who had to move to quarters in
Hagios Mamas on the Bosphorus 109. Political considerations,
based on the strength and importance of the powers behind
the merchants, will have influenced the allocations 110. Venice
was certainly favoured from the Byzantine point of view, since
the location of its settlement provided direct access to the
shops and stores markets via the Makros Embolos.
Secrecy around the Byzantine naval base (see above p. 11)
was no longer a problem, the inexorable silting up of the
Neorion Harbour in the tenth century had possibly led to the
relocation of the arsenal to Sykai; in any case, the fleet was
practically non-existent by the reign of Alexios I 111. Its relative
re-emergence in the twelfth century took place mainly in
Sykai with Latins (from the settlements) providing a welcome
reserve of personnel 112. Finally, it operated – as the fire attack
in 1203/1204 suggests – from the north-west bank (possibly
the Blachernae quarter) of the Golden Horn 113.
Even with the population growth in the tenth century, it
generally remained the case that the former backyard of the
city, profiting from this growth anyway, was always suffi-
cient for foreigners. On the occasion of the violent explosion
of 1182 against Western traders and other residents in the
settlements 114, especially the Venetians, Eusthatios of Thes-
salonica openly spoke (certainly with a polemical undertone)
of the Latin race, which had its separate place on the bank
of the Horn of Byzantium, coinciding with ancient custom 115.
109 Hellmann, Handelsverträge zwischen Kiev und Byzanz. – Shepard, Con­stan­
tinople – Gateway to the North. – Kislinger, Reisen 368-369 with n. 165.
110 Significant is the award of skalai in the flourishing middle section of the
Golden Horn to Germans and French (Jacoby, Venetian Quarter 158-159;
Magdalino, Con­stan­tinople 89), that is (nationals from) states that played an
important role in Manuel’s foreign policy (Magdalino, Empire 41-43. 46-53.
59-66).
111 In the time of Romanos Lekapenos, the arsenal was no longer explicitly as-
sociated with the Neorion (Theophanes Continuatus, Chronographia 391
[Bekker]). In a note in the Patria Kon­stan­tinupoleos II 88 (196 Preger), the
harbour itself is called limne (stagnant water, bog). – In the 13th c., Georgios
Pachymeres V 10 (II 469 Failler) calls the Neorion arsenal really old (palaia
exartysis). – On the decline of the navy, cf. Kislinger, Ruhm 43-52.
112 Lilie, Handel und Politik 614-619. 624-625. 630-633. – Ahrweiler, Byzance et
la mer 282-283. 295. 431-433. – Müller-Wiener, Häfen 12-13.
113 On 1203/1204 Devastatio Con­stan­tinopolitana 90-91 (Hopf). – On the pos-
sible base at the Monastery of Christos Euergetes, see Georgios Pachymeres,
Relationes historicae V 10 (II 469 Failler). – Müller-Wiener, Häfen 8-9. 24. – Cf.
Asutay-Effenberger, Kynegion District, in this volume.
114 Brand, Byzantium 40-43.
115 Eustathios Thessalonikes, Expugnatio 34 (Kyriakidis). – See Jacoby, Quartiers
juifs 181-182. – Magdalino, Con­stan­tinople 99: »Before they were privileged
foreigners, they were just foreigners, and the Golden Horn was their rightful
place«; Rapp, Constantinople and its foreigners 101: »merchants or diplo-
mats, were not normally made to feel ›at home‹, but constantly were re-
minded of their status as outsiders and guests.
 The new arrivals knew better how to use their oppor-
tunities than their hosts could have imagined. Although
entering the Golden Horn required making a sweeping arc
past Chalcedon due to the Bosphorus’ current, the Golden
Horn was itself an enormous natural harbour 116, where large
numbers of ships found anchorage that was easily accessible
from the shore. Cargo could then be unloaded and loaded
at the various skalai (landing stages parallel to the shore
and, presumably, landing stages extending into the water) 117.
The traditional harbour concept was thereby modified and
enlarged. Analogous to the Mese at the harbours of Pro-
pontis, the emboloi here also formed the backbone of the
economic axis 118.
The various commercial settlements grew 119, encouraged
by the diminishing power of the Byzantines towards 1204
to oppose the increasingly demanding wishes of the Italian
maritime powers. It was earlier noted for the period from the
late fourth century to sixth / early seventh century (see above
pp. 9-10) that there was a parallel existence of two economic
centres and harbour zones in one city. This is now repeated
from the end of the eleventh century to the Fourth Crusade
(1203/1204) with the Mese and Propontis harbours, and at
the same time, the riparian zone including land clusters from
Prosphorion to Perama. Constantinople was able to cope
with this once more and even needed it, because of its again
considerably increased population 120.
Again, this mercantile concentration would not be per-
manent. A drastic sequence of events led to the elimination
of an axis (fig. 1) and once again, a significant decline in
population affected sustainability for centuries. The political
background – in the sixth to seventh centuries it had been
constant wars and substantial territorial losses – now formed,
more seriously, the destruction of the empire in the wake of
the conquest of Constantinople in 1204 121. The directly con-
tributory factor of the plague of 542 corresponded to the fires
of 1203/1204 122, added to this was damage from the rioting
of the local mob and looting by the conquerors 123. The artifi-
116 Prokopios, De aedificiis I 5, 13 (IV 29 Haury / Wirth). See Kislinger, Golden
Horn, in this volume.
117 Michael Attaleiates, Historia 199 (Pérez-Martin). Earlier owners were, among
others, monasteries (see above for that of Manuel) or charitable institutions
(such as the Xenon of Isaac II Angelos by the Church of the Forty Martyrs
Niketas Choniates, Historia 445 [van Dieten]. – Acta et diplomata graeca III
16). The proceeds of the skalai (Antoniadis-Bibicou, Douanes 134-135) helped
to meet their expenses.
118 Jacoby, Houses and Urban Layout 271-274. – Magdalino, Maritime Neighbor-
hoods 223-224.
119 Jacoby, Venetian Quarter 156-159.  – Lilie, Handel und Politik 79-81. 101-
102. – Balard, Romanie génoise I 109-112. 179-182.
120 Magdalino, Con­stan­tinople 61-65 n. 28. 45-46. – Schreiner, Costantinopoli
83 estimates 400 000 inhabitants.
121 Queller / Madden, Fourth Crusade. – Carile, Partitio terrarum imperii.
122 Niketas Choniates, Historia 553-554 (van Dieten). – Geoffrey de Villehardouin,
Conquête I § 203. – Madden, Fires.
123 Niketas Choniates, Historia 553-555. 558-559. 570. 647-655 (van Dieten).
124 Van Tricht, Latin renovatio.  – Carile, Storia dell’ impero Latino.  – Jacoby,
Urban Evolution.
125 The lament of Michael Choniates, Epistulae 50, 10 (69-70 Kolovou) from
­Athens, that all goods flow to Constantinople and therefore lack in the prov-
ince, aptly characterises the situation before 1204.
cial Latin empire 124 was too weak (also financially), to remedy
the resulting damage in its urban area of responsibility to
some extent. Its hinterland, which had previously supplied
demand 125, noticeably diminished. In addition, the court and
the upper classes were now lacking as wealthy consumers. In
modern terms, the economy collapsed massively.
Now masters of over three-eighths of the city, the Vene-
tians alone remained a significant economic force126. Because
the trade network of the Italian maritime powers – actively
expanding into the Black Sea region after 1240 127 – stabilised
the entire Rhomania area, Constantinople was able to main-
tain its function as a hub and market, even after the Byzan-
tine reconquista of 1261, naturally adapting to the reduced
internal needs of the city. The total turnover of Byzantine
merchants, at best as a junior partner 128, and large parts of
the retail trade 129 took place in the western settlements and
their neighbourhoods. However, the Mese axis had lost its
commercial importance 130. Of the Propontis harbours, that
of Theodosius was now almost completely silted up 131, the
Kontoskalion 132 (formerly the Harbour of Sophia) was used by
the imperial navy, or what was left of it 133. On two occasions,
the harbour areas of Constantinople once again rose in im-
portance, for naval construction in 1348/1349 134 and during
the final siege of the city in 1453 135. The maritime events
were each focused on the Golden Horn; it was, as ancient
Byzantion had already recognised, the more important and
better location.
Postscript: In the context of the video conference »Columns
of Constantinople«, organised by the Department of Byzan-
tine Archaeology at the University of Freiburg (Germany) on
13 November 2020, Dr Jesko Fildhuth spoke on »Landmarks
or Sea marks? Seeing the Columns of Constantinople«. The
author of the above contribution agrees with Fildhuth that
the towering columns at the Forums of Constantine, Theo-
dosius and Arkadios could have served for the orientation of
incoming ships. However, I do not see this as the original or
126 Jacoby, Economy of Latin Con­stan­tinople. – Jacoby, Venetian Government.
127 Jacoby, Economy of Latin Con­stan­tinople 209-213.
128 Oikonomides, Hommes d’affaires. – Laiou-Thomadakis, Mediterranean Trade
System. – Jacoby, Mediterranean Food and Wine. – Kislinger, Gewerbe.
129 Berger, Ufergegend 154-155. – Kislinger, Lebensmittel 316-318 n. 97 and
99. – Mundell Mango, Commercial Map 205-206.
130 Concerning »commercial buildings« in this area, we know of only two baker-
ies in the »Old Forum« (that of Constantine) and wine taverns in the harbour
area: Kidonopoulos, Bauten 203-204. 211-212. – Kislinger, Lebensmittel 310
n. 47.
131 Berger, Langa Bostani 471-472.
132 Makris, Studien 176-184. 288-290.  – Müller-Wiener, Häfen 26-28.  – Cf.
Heher, Harbour of Julian, in this volume.
133 Ahrweiler, Byzance et la mer 375-378. 433. – Georgios Pachymeres, Rela-
tiones historicae V 10 (II 469 Failler) notes bitterly that the Golden Horn must
now be shared with the ships of the enemies.
134 See Preiser-Kapeller, Haptaskalon, in this volume. – Nicol, Last Centuries 228-
233. – Nicol, Reluctant Emperor 96-99.
135 Runciman, Fall of Con­stan­tinople 100-111. – Pertusi, Caduta di Costantino­
poli.
On Better and Worse Sites  |  Ewald Kislinger 17
primary purpose for their construction. This is, for that mat-
ter, partly impossible due to the time difference in building
forums and harbours (Column of Constantine, Harbour of
Julian). However, the positioning of the forums and columns
Summary / Zusammenfassung
On Better and Worse Sites: The Changing Importance
of the Harbours of Constantinople
Alternating phases of growth and demographic decline in
Constantinople led to the displacement of the central traffic
axes and their associated harbours on two occasions. An-
cient Byzantion was orientated towards the Golden Horn,
where its two harbours of Neorion and Prosphorion were
located. When Constantinople became the imperial capital
after 330, the population increased and the urban area was
extended, with new harbours, named after Julian and Theo-
dosius, established on the southern shore facing the Sea of
Marmara. With the Mese, the main street as the spine run-
ning  ­east-west, an additional traffic axis was created. As the
population decreased after the sixth / seventh century, partly
as a result of plague outbreaks, the Mese alone remained.
The Golden Horn was now off-centre and became the base
for the imperial navy. It was only the revival from the tenth
century onwards, which also saw the return of mercantile
activity to the area. The commercial settlements, which Byz-
antium had to cede to the Italian maritime powers, were
granted to them on the Golden Horn, apparently in the false
assumption of keeping them well away from the commercial
centre. The opposite occurred, the entire coastal strip at the
mouth of the sea gained in importance: by the twelfth cen-
tury, it was of equal rank and after 1204 became the new
centre of Constantinople's maritime economy.
18 On Better and Worse Sites  |  Ewald Kislinger
along the Mese axis undoubtedly underlines their significance
for the growing importance of the harbours on the Propontis
at that time. (I would like to thank my colleague Fildhuth for
making his lecture manuscript available to me.)
Von schlechteren und besseren Lagen. Häfen zu Kon­
stan­tinopel im Wandel ihrer Bedeutung
Die abwechselnden Phasen von Wachstum und demographi-
scher Schrumpfung resultierten zu Kon­stan­tinopel in einer
zweimaligen Verlagerung der zentralen Verkehrsachse und
der zugehörigen Häfen. Das antike Byzantion war auf das
Goldene Horn hin orientiert, dort lagen seine beiden Häfen,
Neorion und Prosphorion. Als Kon­stan­tinopel nach 330 zur
Reichshauptstadt avancierte, wuchs daraufhin die Bevölke-
rung, das Stadtareal wurde erweitert, neue Häfen, benannt
nach Julian und Theodosios, entstanden an der Südküste
am Marmarameer. Mit der Mese, der in Ost-West-Richtung
verlaufenden Hauptstraße als Rückgrat war somit eine zu-
sätzliche Verkehrsachse entstanden. Sie allein verblieb, als
die Bevölkerung, unter anderem durch die Pestwellen be-
dingt, vom 6./7. Jahrhundert an schrumpfte; das jetzt im
Abseits liegende Goldene Horn wurde zum Stützpunkt der
kaiserlichen Marine. Erst der neuerliche Aufschwung ab dem
10. Jahrhundert brachte auch merkantile Aktivitäten dort-
hin zurück. Die Handelsniederlassungen, welche Byzanz den
italischen Seemächten einzuräumen hatte, wurden ihnen
am Goldenen Horn zugewiesen, offenbar in der falschen
Annahme, sie derart abseits des kommerziellen Zentrums
zu halten. Das Gegenteil trat ein, die ganze Ufergegend am
Meeresarm gewann an Bedeutung, wurde schon im 12. Jahr-
hundert gleichrangig und nach 1204 wiederum zum neuen
wirtschaftlich-maritimen Zentrum Kon­stan­tinopels.
Arne Effenberger
Constantinople / İstanbul:
The Early Pictorial Sources
The illustrations included in the individual chapters on the
harbours have a varying degree of testimonial value, both in
terms of the state of the city of Constantinople / İstanbul and
in terms of the representation of the harbours and landing
places. This applies especially to the early city views from the
fifteenth and sixteenth centuries. Therefore, in the following,
individual vedute will be considered more closely with regard
to their time of origin and their possible sources.
Cristoforo Buondelmonti
The view of Constantinople connected with the name of Cris-
toforo Buondelmonti ranks first. It is preserved in numerous
copies of the Liber insularum Archipelagi (hereinafter referred
to as LIA) (fig. 1). Cristoforo Buondelmonti (c. 1380/1385 to
around 1431) travelled extensively through the islands of the
archipelago, visiting the Ionian west coast, Constantinople
and Mount Athos since the year 1414 1. He collected the re-
sults of his geographical and archaeological research in two
works: the Descriptio insulae Cretae (hereinafter referred to
as DIC) and the LIA. In 1417, he dedicated a first version of
the DIC to his mentor Niccoli Niccolò, as the acrostic in the
chapter beginnings shows 2. He sent a new version of the
LIA to his patron Cardinal Giordano Orsini († 1438) in 1420 3.
Again, he used an acrostic indicating the cardinal as described
in the prologue: In quibus dum rubeas ennumerabis ipsarum
litteras, nomen meique tui et quo in locoque tempore pre-
feceram opus manifestabis (»If you follow the red initial let-
ters of the chapters subsequently, you will be able to find
the letters of my and your name as well as where and when
I wrote my work«) 4. The first letters of the 82 chapters make
up the words: CRISTOFORVS BONDELMONT DE FLORENCIA
1 On the person, see Weiss, Buondelmonti.  – For Buondelmonti’s work, see
Barsanti, Costantinopoli.  – Vagnon, Cartographie 273-304.  – Chatzidakis,
Ciriaco d’Ancona 41-48.
2 Barsanti, Costantinopoli 83. 102. 111-127. – Ragone, Buondelmonti 193-194
and n. 51.
3 This is probably the long version A, which has survived at least in three
manuscripts: 1: Milan, Biblioteca Ambrosiana, Ms. A.219 inf.: Gerola, Vedute
270-279 (Constantinople chapter). – 2: Ravenna, Biblioteca Civica Classense, Ms.
Lat. 308: Gerola, Vedute 270-279 (Constantinople chapter).  – 3: Greenwich,
National Maritime Museum, Ms. p. 20.  – See Barsanti, Costantinopoli 160.  –
Ragone, Buondelmonti 195. 200.
4 See Bayer, Transkription 8 § 2 (14). – See Effenberger, Illustrationen, pl. 1. Here
ends the year number MCCCCXX or possibly MCCCCCC. On the possible
solutions of the chapters 81 and 82 beginning with X, see Ragone, Buondelmonti
199 n. 70.
In: Falko Daim  ·  Ewald Kislinger (eds), The Byzantine Harbours of Constantinople. Byzanz zwischen Orient und Okzident 24 (Mainz 2021).
DOI: https://doi.org/10.11588/propylaeum.911.c12064
PRESBITER HVNC MISIT CARDINALI IORDANO DE VRSINIS
MCCCCXX (»The Presbyter Cristoforus Bondelmont from
Florence sent this to Cardinal Jordanus Ursinus in 1420«) 5.
Where he wrote the work, however, does not emerge from
the acrostic. The sixty-seventh letter D of DE VRSINIS intro-
duces the chapter on Constantinople: Devenio ad Lesam,
nunc Constantinopolim urbem .... Buondelmonti concludes
the epilogue (chapter 82) with two remarks: that he had al-
ready sent the cardinal an initial version on the Cyclades, and
that now, after more careful investigations, he wanted to give
him a second, more elaborate version (secundam copiosiorem
etiam tibi desculptionem uoli destinar) 6. That this copy con-
tained the 79 images of the islands and cities can be found
in chapter 2 of the prologue: »To capture everything even
better: the mountains are represented in black, the plains in
white and the waters in green« 7. There was no doubt that in
the pictures of the islands and cities the places mentioned in
the text were also indicated and provided with inscriptions 8.
Buondelmonti must, therefore, have been in Constantinople
for the first time before 1420 and had ample opportunity to
explore the city and make a detailed vedute.
The first copies of Buondelmonti‘s works were already
produced during his lifetime; the majority of the copies date
from the second half of the fifteenth century. Hitherto, the
impressive number of more than 70 manuscript copies of
the LIA has been validated, including those in Greek, Italian
and English translation 9. Several manuscripts are dated by
copyist entries or can be narrowed down to a time frame by
the names of their owners, by watermarks in the paper, occa-
sionally by textual or internal indications, which at the same
time demonstrate the rapid spread of the LIA in Europe (Italy,
France, Flanders and England) 10. However, the textual tradi-
tion of the LIA is complicated, since in addition to the long
  5 Barsanti, Costantinopoli 85 n. 7. – Ragone, Buondelmonti 195.
  6 Bayer, Transkription 59 § 82.
  7 Bayer, Transkription 8 § 2 (15).
  8 A list of the cardinal’s surviving books, published in 1786, lists a Liber insula-
rum egei pelagi and a Liber insularum archipelagi et figuratus, see Effenberger,
Illustrationen 14 n. 28 and 20 (evidence).
  9 List of 60 copies: Luttrell, Halikarnassos 193-194. – Ragone, Buondelmonti 202-
203 meanwhile assumes »circa settantacinque« copies.
10 On dated and datable copies of the LIA and on previous owners, see Barsanti,
Costantinopoli 86-91. – Ragone, Buondelmonti 181 n. 12.
Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger 19
Fig. 1  View of Constantinople and Pera, Cristoforo Buondelmonti, Liber insularum archipelagi. – (Venice, Biblioteca Nazionale Marciana, MS. Lat. XIV.45 (= 4595), fol. 123r).

20 Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger

version A, an abbreviated version B (the majority of surviving
copies) and a short version C can be distinguished 11. The
only printed edition of Gabriel R. L. de Sinner (1824), which
belongs to version B, was compiled from three manuscripts in
Paris (Bibliothèque nationale de France, MS. Lat. 4823, 4824
and 4825) 12. The text of the Düsseldorf manuscript, which
also follows version B, is available in a transcribed, translated
and annotated edition 13.
An unresolved problem is the acrostic’s deviation of the
date in the incipit that is found in both version A and in
copies of version B: Incipit Liber insularum archipelagi editus
per presbyterum Christophorum de Bondelmontibus de Flor-
entia, quem misit de civitate Rhodi Romam Domino Jordano
Cardinali de Ursinis anno Domini MCCCCXXII (»Book about
the islands of the archipelago written by the presbyter Chris-
tophorus Bondelmonti from Florence, which he sent from the
city of Rhodes to Rome to Cardinal Jordanus Ursinus in the
year of the Lord 1422«) 14. A manuscript in the Vatican, made
between 1422 and 1435 on Rhodes by Onofrio da Penna,
secretary of King Ladislao d’Angiò-Durazzo (1386-1414) and
Queen Giovanna II of Naples (1414-1435), contains the short
version C of the LIA (without the view of Constantinople), as
well as a new elaboration of the DIC, with Buondelmonti‘s
colophon on fol. 50v 15: Scripsi hunc librum figuramque insule
in urbe Constantinopoli, die .xviij°. mensis Ianuarii .m°cccc°.
xxij°. (»This book and the depiction of the island [Crete],
I completed in the city of Constantinople on 18 January
1422«). This entry testifies to Buondelmonti’s second stay
in Constantinople at the end of 1421 or the beginning of
1422 16. The following insertion is found in the Constantinople
chapter of Chigianus (see n. 15) and another manuscript copy
of version C made before 1453 17: ideo quam brevius potui
hic de ruynis eius scripsi, licet in membrana maxima Bittoldo
ducis Russie miserim, ad videndum suis omnibus extra atque
infra attinentiis (»therefore, I want to write here only briefly
of their remnants, especially since I have sent a large sheet to
Witold, Duke of Russia, to show what all is contained outside
and within [the city]«). Witold (Biteldos, Vitovt, Vytautas),
Grand Duke of Lithuania (r. 1392-1430) 18, was the father-
in-law of the Muscovite Grand Prince Vasily  I Dmitriyevich
11 Barsanti, Costantinopoli 160-164. – Ragone, Buondelmonti 193 n. 50 (list of
the surviving manuscripts of version C).
12 Sinner, Buondelmonti. – See Garand, Tradition 69-76. – Buondelmonti’s LIA in
Greek translation in the manuscript İstanbul, Topkapı Sarayı Kütüphanesi, Ms.
Seragliensis Graecus 24.
13 Beyer, Transkription 50-53.
14 On this unsolved contradiction, see Barsanti, Costantinopoli 161. – Ragone,
Buondelmonti 195-196.
15 Città del Vaticano, Biblioteca Apostolica Vaticana, Ms. Chigi F. IV.74: Barsanti,
Costantinopoli 86-87.  – Ragone, Buondelmonti 206-208. The subscriptio is
still present in the following copies of the LIA and DIC: 1: Paris, Bibliothèque
nationale de France, Cartes et Plans, Ms. Rés. Ge. FF 9351: Luttrell, Halikarnas-
sos 194 no. 31. – 2: Padua, Biblioteca Universitaria, Ms. 1606: Luttrell, Halikar-
nassos 194 no. 30. – 3: Città del Vaticano, Biblioteca Apostolica Vaticana, Ms.
Ross. 703 (according to the entry on fol. 24r copied by Bartolomeo de Columnis
of Chios, in which the incipit on fol. 1r gives the year 1425). – On the dating of
the individual versions, see Ragone, Buondelmonti 193-194.
16 The short version C of the LIA contains neither the acrostic nor the incipit of
1422.
Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
(r. 1389-1425). His daughter Anna  – Witold‘s granddaugh-
ter  – probably came as an infant bride to Constantinople
in 1411 and was married to the future Byzantine Emperor
John  VIII Palaiologos (r. 1425-1448) in 1414. The Russian
deacon Zosima had been part of the bridal legation and in
1419/1420 and 1421/1422 again visited Constantinople. He
also visited the Princess’s grave at Lips Monastery; she had
died of the plague in 1417 19.
However, the date of composition of the large-sized view
of Constantinople that was sent to Witold is difficult to de-
termine. How and by which intermediary Buondelmonti was
commissioned to produce the membrana maxima, remains a
mystery despite many efforts to illuminate the historical con-
text 20. It would seem logical that Buondelmonti first explored
the city and its surroundings with official permission before
producing the membrana maxima and only afterwards cre-
ated the standard vedute added to the LIA, the secunda co-
piosior of 1420. Thus, the membrana maxima would already
have been produced during his first stay in Constantinople
(before 1420). In any case, it can be assumed that the dis-
semination of the Constantinople veduta, which began soon
after 1420, can only have been stimulated by one or more
early copies of Buondelmonti‘s own hand.
The Constantinople veduta is preserved in numerous cop-
ies of the LIA, but the individual reproductions are of different
quality. A systematic »Kopienkritik« (copy criticism) on the
basis of all surviving versions is still pending 21. Although the
Buondelmonti-type vedute have a largely consistent standard
of buildings and monuments, there are divergences with
respect to those mentioned and illustrated in the three LIA
versions 22. In any case, all copies made after 1453 are pre-
dominantly anachronistic. Monuments that no longer existed
before or after 1453, yet were still reproduced, include the
Church of Blachernae (burnt down in 1434), the equestrian
statue on the Column of Justinian, the cross on the Column
of Constantine and the the kneeling emperor on the columna
virginea, the Column of Michael VIII Palaiologos 23. Most of
the churches inscribed with their names were either no longer
extant (e. g., the Church of the Holy Apostles demolished
in 1462) or converted into mosques, but on many copies
17 Venice, Biblioteca Marciana, Ms. Lat. X.215 (= coll. 3773), fol. 44r (previously
owned by Francesco Barbaro, 1390-1454). The Constantinople chapter printed
in Gerola, Vedute 270-279. – See Barsanti, Costantinopoli 168-169. – Ragone,
Buondelmonti 205-208.
18 PLP 2 (1977) 68 no. 2708. – Mickūnaitė, Making a Great Ruler.
19 Majeska, Russian Travelers 170. 188-189. 309. 311-312 § 34. – On Anna see
PLP 1 (1976) 94 no. 1003.
20 Ragone, Membrana maxima. – Ragone, Buondelmonti 205-217.
21 On this problem, see Barsanti, Costantinopoli 164-197.
22 Effenberger, Illustrationen 18 n. 19 are first, those mentioned, but illustrated
without inscription; second, those mentioned, but not illustrated, and third,
those not mentioned, but with records of illustrated buildings and monuments.
The list is in need of revision, as all Constantinople copies would have to be
considered. With regard to the Monastery of Christ Pantepoptes, see n. 25. –
Gerola, Vedute 266-269, has put together the inscriptions for 10 versions in a
synoptic table. In addition, however, there are further inscriptions on vedute
not used by him.
23 Asutay-Effenberger / Effenberger, Columna virginea.
21
they still bear a crowning cross. Only in a few cases, reliably
datable elements are included by later copyists, such as the
representation of the Yedikule Fortress (built 1457/1458) on
a veduta in Paris 24. Here you will also find the only labelled
reproduction of the Monastery of Christ Pantepoptes known
to me 25.
Likewise, the harbours depicted have not been updated
to include changes made after 1453. Referred to in the LIA
are the olim portus velanga 26, contescali vel arsana 27 and
portulus imperatoris 28; depicted on several copies are the
former harbour by the Mermerkule (Marble Tower; portus sed
destructus preceptu turchorum), the Harbour of Theodosius
(Vlanga), Bucoleon (portus olim palatii imperatoris) and Kon-
toskalion (Receptaculum dictum contiscali) 29. The harbour by
the Mermerkule is often marked by a south-westerly pier and
on representations where the mole is missing the impression
is created that the ditch was in direct connection with or ran
into the Sea of Marmara 30. Although it is not possible to de-
cide whether the harbour depiction is a more recent element
and is based on local historical memory referring to the events
of 1390/1391 31, but I do not consider tracing it back to one
of Buondelmonti’s replicas to be out of the question.
In most of the Buondelmonti views of Constantinople
(fig. 1), the river Lycos begins within the city (incorrect), runs
north of the Column of Theodosius (incorrect) 32 or between
this and the Column of Arcadius (correct) 33 and ends in the
Harbour of Kontoskalion (incorrect). The Lycos flows correctly
into the bay of Vlanga in only a few copies from the second
half of the fifteenth century 34. In contrast, a body of water
rises at the Church of Peribleptos and pours into the harbour
bay of Vlanga in numerous verdute 35. Both errors could be
traced back to Buondelmonti‘s original version, since it was
apparently forgotten where the extensively canalised Lycos
ended. However, it cannot be ruled out that the sewers of the
Kainoupolis district flowed into the still open bay of the Kon-
toskalion. This may also explain the inexorable silting up of
the inner harbour basin, which led to its disuse in the middle
of the sixteenth century. The anonymous Armenian pilgrim
24 Paris, Bibliothèque nationale de France, Ms. Nouv. Aqu. Lat. 2383, fol. 34v: Ef-
fenberger, Polichnion, with fig. 1. – Ganschou, «La Tour d’Irène» 159-200 fig. 4.
25 Buondelmonti mentioned the cisterna panda pophti (Bayer, Transkription 91
§ 40), whereby is meant the Çukurbostanı (Cistern of Aspar) by the Yavuz Selim
Camii. As a result, locating the monastery on the sixth hill, proposed by Mango,
Monastery of Christos Pantepoptes, gains further support.
26 Beyer, Transkription 50 (8): ollim portus velanga.  – On the harbour, Külzer,
Harbour of Theodosius, in this volume.
27 Ibid. 50 (9): condescali uel arsana. – On the harbour, see Heher, Harbour of
Julian, in this volume.
28 Ibid. 50 (11): portulus imperatoris. – On the harbour, see Heher, Harbour of the
Boukoleon, in this volume.
29 The different name variants on six copies in Gerola, Vedute 269-269.
30 On the scarce information on the Brachialion, which sealed off the ditch against
the sea, see Simeonov, Brachialion, in this volume.
31 On this, see Majeska, Travelers 100-104 (account of Ignaty of Smolensk) 408-
415 §§ 81-87).
32 Venice, Biblioteca Marciana, MS. Lat. XIV.45 (=4595): Barsanti, Costantinopoli
fig. 61; see also figs 83-84. 86-87. 90-91.
33 Barsanti, Costantinopoli figs 48. 60. 64. 72. 75. 77-78. 100.
34 Ravenna, Biblioteca Civica Classense, Ms. Lat. 308, fol. 58v: Barsanti, Costan-
tinopoli 100 fig. 64. – Città del Vaticano, Biblioteca Apostolica Vaticana, Ms.
22 Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
who visited Constantinople in 1434 is the only one who
mentioned the sweet water that sprang from under the Mon-
astery of Peribleptos 36. That the water actually flowed down
the valley as an open brook before the spring in the Hagiasma
of the Armenian church Surp Kevork was set 37, must surely
be assumed. On the contrary, although the above-mentioned
Parisian copy (see n. 24) shows the Church of Peribleptos, but
indicates the mouth of the brook just below the Stoudios
Basilica and has also moved the harbour of Vlanga here with
the indicated mole.
The large-format view in the Düsseldorf version of the
LIA, which can be dated to 1485-1490 on the basis of the
watermarks and image-related evidence (fig. 2) 38 differs fun-
damentally from the traditional Constantinople depictions.
This view is again based on accurate local knowledge of
Mehmed II‘s city, supplemented by a few updates from the
time of Bayezıd II 39. Besides Constantinople, Eyüp, Pera and
a part of Üsküdar (Scutari), it also covers the territory on both
sides of the Bosphorus to the coast of the Black Sea. This
representation also contains numerous »deliberate« anachro-
nisms 40 and, as an innovation, the Valens aqueduct, which is
missing on all the Buondelmonti vedute 41. However, the mina-
rets of the Fatih Camii have been omitted. Vlanga (inscription:
Vlanga locus aquosus) is only depicted here as a completely
walled-in area with eleven towers (fig. 3) 42. The views of the
Buondelmonti type normally show a southern pier and usually
mark the former harbour basin as already silted up (fig. 1) 43.
The Kontoskalion – here already operating as a galley harbour
(Kadırgalimanı) and shipyard – is subdivided into outer and
inner harbour basins (inscription: mare) and closed with an
iron gate, with five ship sheds (inscription: darsinale) includ-
ing warships of the Kadırga type lying within 44 and a galley
moored on the inner bank (fig. 4) 45. Another arsenal, proba-
bly built under Mehmed II, with two ship sheds and a landing
stage in front of a gate of the Serail or the Marmara seawall
(Değirmenkapı?), with the inscription darsinale regiu(m), is
located in the area of the former Mangana (fig. 5) 46. A regular
pier is located in front of the cannon foundry (Tophane), two
Ross. 702, fol. 32v (watermark c. 1475/1478): Barsanti, Costantinopoli 182
fig. 78.  – London, British Library, Ms. Arundel 93, fol. 155r (colophon from
1485): ibid. 196. 207 fig. 75.
35 See Barsanti, Costantinopoli figs 48. 61. 64. 77. 83-84. 87. 91; only the water
course without connection to the Church of Peribleptos: fig. 72.
36 Brock, Description 88. 98-99. – The Peribleptos Monastery had a good natural
spring, as the Turkish name Sulu Manastır (»monastery with water«) indicates.
The source was once under the altar of the neighbouring Church of St Stephen
of Aurelianus, see Müller-Wiener, Bildlexikon 200.
37 Atzemoglu, T‘agiasmata 51-52.
38 Effenberger, Illustrationen 67-68.
39 The Türbe of Mehmed II, which was built after his death (1481).
40 Effenberger, Illustrationen 67.
41 Effenberger, Illustrationen 48 no. 35 fig. 32.
42 Effenberger, Illustrationen 31-33 no. 13 fig. 32. – See Külzer, Harbour of The-
odosius, in this volume.
43 For example, see Barsanti, Costantinopoli figs 48. 60-61. 64. 70-75. 77-81.
83-84. 87. 91. 94. 97.
44 On the types, see Bostan, Osmanlı Bahriye Teşkilâtı 85.
45 Effenberger, Illustrationen 29-31 no. 12 fig. 32. – See Heher, Harbour of Julian,
in this volume.
46 Effenberger, Illustrationen 26 fig. 32 no. 7.
 Fig. 2  View of Constantinople, Pera and
the upper Bosphorus by Cristoforo Buon-
delmonti, Liber insularum archipelagi.
Düsseldorf, Universitäts- und Landesbib-
liothek, MS. G 13, fol. 54r, c. 1485-1490
(the manuscript is on loan from the City
of Düsseldorf to the Universitäts- und
Landesbibliothek of Düsseldorf).

Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger 23


more and a galley on the shore of Üsküdar (Scutari). It also
becomes clear that the entire shore zone of the Golden Horn
of Galata served as a harbour (fig. 1).
Giovanni Andrea Vavassore
The 1530 woodcut by Giovanni Andrea Vavassore with the
cityscape of Constantinople (fig. 6) 47 is attributed to a proto-
type that cannot have been made earlier than 1478/1479 48.
The terminus post quem is given by the representation of
the wall of the Topkapı Sarayı, which was completed in
47 I know of only four copies of the woodcut: 1: Nürnberg, Germanisches Na-
tionalmuseum, SP 8190, Kapsel 1102 (52.6 cm × 36.8 cm): Fauser, Repertorium
379 no. 6817. – Berger, Vavassore 329-355 fig. 1. – Manners, Image 91-94
fig. 8.  – 2: Staatsbibliothek Bamberg, Sign. IV C 44 (kindly pointed out by
Bernhard Schemel): Effenberger, Illustrationen fig. 5. – 3: Lyon, Bibliothèque
Municipale. – 4: California State Library, Coll. Fullerton.
48 Analysis of individual buildings: Berger, Vavassore 339-355.
49 It can be ruled out that in the Vavassore woodcut buildings from the time of
Bayezıd II (1481-1512) are already included, see Necipoğlu, Visual Cosmopoli-
tanism 70 n. 125.
50 Stichel, Coliseo 448-459.
24 Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
Fig. 3  Detail from fig. 2 showing the
Harbour of Theodosius, the Church of
St Mary Peribleptos and the Basilica of
St John Stoudios. – (Düsseldorf, Uni-
versitäts- und Landesbibliothek, MS. G
13, fol. 54r, c. 1485-1490).
1478/1479 49. The Vavassore view has some inconsistencies
in the city centre (the possible doubling of the sphendone
of the Hippodrome, wrong or transposed inscriptions) and
is probably based on an intermediate draft that would have
already contained these errors 50. Dependant on the latter
are probably also the Constantinople veduta in Sebastian
Münster‘s Cosmographia of 1550 51, the large, double-sided
engraving by Giulino Ballino of 1567 52 and the Constantino-
ple view in the Civitates of Georg Braun and Franz Hogen-
berg (1572) 53. They all contain deviations and innovations in
individual cases compared to Vavassore‘s view. The common
intermediary draft was probably based on a large-format
51 Sebastian Münster, Cosmographia, Basel 1550 (simultaneous German and
Latin editions) 940-941 (as well in all reprints from 1552-1628). – See Fauser,
Repertorium LXI, 379 no. 6818 (woodcut by David Kandel). – Effenberger, Il-
lustrationen fig. 6 (after the Latin edition in the Staatsbibliothek Bamberg, Sign.
J. H. Geogr. f. 2).
52 Giulino Ballino, Disegni, without pagination. – See Fauser, Repertorium XXII,
379 no. 6820 (unsigned; stamped 1567). – Effenberger, Illustrationen fig. 7.
53 Braun / Hogenberg, Civitatis orbis terrarum, Cologne 1572, vol. 1 pl. 51 (=
pl. 52 in the German edition). – See Fauser, Repertorium XXXIV-XXXV, 379
Nr. 6824 (Kupferstich). – Berger, Vavassore 329-331 fig. 2.
Fig. 4  Detail from fig. 2 showing the harbour and arsenal of Kadırgalimanı and
the stables. – (Düsseldorf, Universitäts- und Landesbibliothek, MS. G 13, fol. 54r,
c. 1485-1490).

Fig. 6  View of Constantinople, Pera-Galata and the Asiatic coast by Andrea Fig. 5  Detail from fig. 2 showing the Sultan’s arsenal and landing stage at Top-
Vavassore. Woodcut, c. 1530, after an original from 1478/1479-1481. – (Bam- kapı Sarayı. – (Düsseldorf, Universitäts- und Landesbibliothek, MS. G 13, fol. 54r,
berg, Staatsbibliothek, IV C 44). c. 1485-1490).

Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger 25


Fig. 7  Detail from fig. 6 showing the area between Yedikule and Vlanga. –
(Bamberg, Staatsbibliothek, IV C 44).
view of Constantinople consisting of several single sheets.
There are some candidates for this. The inventory of the
bequests of the Florentine cartographer Francesco Rosselli
(† after 1508) lists among the printing plates different views
of Constantinople: Gostantinopoli in 6 pezzi; gostantinopoli
in tela cholorita inn istampa del pupillo; francia chon parte
gostantinopoli, in mezzo goglie comume 54. Also in the ex-
tensive printmaking collection of Johannes Columbus (1488-
1539) in Seville was a five-page woodcut by the Florentine
Lucantonio degli Umberti (Venice, c. 1510-1529) 55 and an
eight-leaf view of Constantinople. On the latter, Peter Barber
remarked: »Its composition, as described in the inventory,
leaves little doubt that it was the model for the single-sheet
woodcut of about 1520 by Giovan Andrea Vavassore […] It
is possible that the print may have some relationship with or
may even be the six-sheet printed view of Constantinople by
54 Del Badia, Bottega 24-30. – Hind, Early Italian Engraving 1, 304. 305-306 (inv.
III no. 60; inv. I no. 4; inv. III no. 47). Manners, Image 94 and Stichel, Coliseo
454. 459, however, judges tracing back Vavassore’s woodcut to Rosselli with
scepticism, although the latter basically insists on a large-sized six-part proto-
type.
55 Barber, Maps vol. 1, 261 n. 77; vol. 2, 569, inv. no. 3159. – See also Hind, Early
Italian Engraving 1, 211-214.
56 Barber, Maps vol. 1, 255; vol. 2, 573, inv. no. 3178. The author does not cite
the works of Berger, Vavassore and Stichel, Coliseo, nor is he aware that an-
other copy of Vavassore’s woodcut exists in the Staatsbibliothek of Bamberg.
57 Necipoğlu, Visual Cosmopolitanism 70 n. 125, is also sceptical.
58 Necipoğlu, Visual Cosmopolitanism 27.
59 On the depiction of the three kiosks in the Topkapı Sarayı, see Necipoğlu, Visual
Cosmopolitanism 27.
60 The latest date for the creation of the original – 1490 – was long argued on
the grounds that the church in Vavassore’s woodcut named as S.  Luca Eu-
angelista was the Nea Ekklesia or the Güngörmez kilisesi, which served as a
baruthane (powder magazine) and on 12 July 1490 was destroyed by lightning,
see Mango, Brazen House 180. – Mango, Développement 9 n. 9. – Mango,
Nea Ekklesia: ODB II (1991) 1446; followed by Effenberger, Illustrationen 19.
The Ottoman written sources for the thunderstorm on 12 July 1490 in Konyalı,
26 Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
Rosselli with two sheets of decorative material added« 56. I
think that this is not very convincing, since there is mention
of six (Rosselli) and once of eight single sheets (Columbus)
and the description in the inventory gives too few references
to individual representational elements 57.
However, regardless of this, the prototype and the postu-
lated intermediate draft already contained numerous ships of
different types (galleons, galleys), some of which carry flags
with the Byzantine double eagle, Venetian lion, Ottoman
crescent and Genoese cross. In this respect, Gülru Necipoğlu
referred to the treaty between Mehmed II and Venice of 1479
and considered the creation of the Vavassore prototype soon
after that date as possible 58. Whoever created the prototype,
the question also arises how he was able to explore the city
and in particular the two serails within its walls in such detail,
which could certainly only have happened with the permis-
sion of Mehmed 59. This suggests a dating of the original
before 1481 60.
Vavassore’s veduta shows the city from a bird’s eye view
like Buondelmonti’s depiction (fig. 1), but from the Asian side.
Noticeable distortions can be found in the southern part of
the city (top left). The church with the river, i. e., the Church
of Peribleptos, is wrongly labelled as St  Lazarus (fig. 7) 61.
Against the course of the coastline, the Vlanga (inscription:
Lauulaca) is stylised as a rectangular walled area with six
towers and is marked with three bushes in accordance with
its function as a bostan (garden). To the west is the fenced-in
area of Küçük Langa Bostanı. The watercourse is no longer
shown as in the Düsseldorf view. In the case of Kadırgalimanı,
Vavassore offers an irregularly immured area for the inner part
of the Tersane with the portal-like iron gate (see p. 63, fig. 8).
The five ship sheds with far too small entrances and the single
galley in the inner basin are apparently rotated by 180° for
better clarity. The above-mentioned Vavassore-type vedute
have freely spun out the interior (see pp. 64-65, figs 10-12).
The former imperial harbour and the façade of the Palace of
Bucoleon are missing. Only the inscription porta leona de la
riua next to a gate in the sea wall refers to the portal with
İstanbul Sarayları 18-19. The church appears to be much more detailed and
still intact, but without an inscription on the engraving in Panvinio, De ludis
circensibus 61 pl. R. Mango, L’Euripe de l’hippodrome 182 n. 5, dated the
original to after 1491, because the Firuz Ağa Camii, built 1491, was already
depicted. The Nea Ekklesia possibly existed as a ruin for a longer period. The
Bolognese scholar Luigi Fernando Marsili (1658-1730), who, as a young man,
lived in İstanbul in 1679/1680 in the service of the Venetian Bailo, reported: giù
verso le stalle v’é una chiesa di stile greco, cinta con pilastri di muro, avendo
tre capelle e la porta. Di fuori ha aspetto buonissimo, e nella volta si scoprono
anche alcune vestigie di mosaico (»Down in the direction of the stables, there
is a Greek-style church, surrounded by pilasters on the wall, which has three
chapels [apses] and a door. From the outside it has a very beautiful appearance,
and on the vaulting one spot even some remnants of mosaic«), see Paribeni,
Chiesa antica greca nel serraglio posta 318 (Bologna, Biblioteca Universitaria,
Fondo Marsili, MS. 51, c. 356v). By stalle he means the stables of Mehmed II,
which are shown for the first time in the Düsseldorf İstanbul view with the
inscription stabula regis, see Effenberger, Illustrationen 28 no. 10 fig. 32 (see
my fig. 4). The church was, therefore, west of the stables in the former palace
area. – See now Effenberger, St. Grovus.
61 Berger, Vavassore 349 no. 35.
Fig. 8  Depiction of the storm of 12 July 1490 in Hartmann Schedel’s Liber Chronicum. – (From Hartmann Schedel, Liber Chronicarum (1493), fol. 257r, com-
plete edition, coloured, 1493, introduction and commentary by S. Füssel [Cologne 2001]).

Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger 27


Fig. 9  Detail from fig. 2 showing the Hagia Sophia and the Column of Justinian
with equestrian statue. – (Düsseldorf, Universitäts- und Landesbibliothek, Ms. G
13, fol. 54r, c. 1485-1490).
spolia below the (demolished) western palace front (see p. 69,
fig. 2; p. 76, fig. 19) 62. The Golden Horn was probably (mis)
understood by the Venetian Vavassore alone as the »Canal
Grande« and filled with eight gondolas that serve as traghetti
(fig. 6).
Hartmann Schedel
The 1493 Liber chronicarum, the world chronicle of the
Nuremberg physician and Humanist Hartmann Schedel
(1440-1514), contains three views of Constantinople 63: a
large-scale representation of the entire urban area including
Pera on two pages on folios 129r-130v (see p. 44, fig. 5); a
reduced reproduction of this representation on folio 249r;
and a section of the eastern city area on folio 257r (fig. 8),
62 Effenberger, Illustrationen 28-29. – On the Bukoleon Palace, see Heher, Har-
bour of the Bukoleon, in this volume.
63 Fauser, Repertorium LXV-LXVI, 378-379 nos 6810-6816 (all prints from 1493-
1497).
64 Berger / Bardill, Representations of Constantinople 2-14 see the originals in
Buondelmonti’s vedute and in the protoype of Vavassore’s view. Not all at-
tempts at identification are convincing. The towering »pole« on a two-level
substructure between the columns of Justinian and Theodosius certainly stands
for the Obelisk of Theodosius in the Hippodrome. The completely immured and
tree-covered district with the gate from which a brook gushes out, is Vlanga (as
Külzer in this volume concurs), a combination of the Buondelmonti and Vavas-
sore type. The immured district on the Golden Horn does not indicate the outer
wall of Leo V, but the Phanarion. Both are illustrated only on the Düsseldorf
view (the latter with inscription fanarium), see Effenberger, Illustrationen 39-40
nos 21-22 fig. 32. The Phanarion by itself only appears in Vavassore’s veduta.
This would indeed speak in favour of the fact that an intermediate template or
a copy of the Vavassore type was known to the illustrator.
65 Analysis: Berger / Bardill, Representations of Constantinople 15-24.
28 Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
on which the devastating storm of 12 July 1490 is rep-
resented (see n. 60 above). It must be the case that the
woodcutter Wilhelm Pleydenwurff produced images from
various models from before 1493 for all three illustrations 64.
This issue arises, above all, for the compilation of build-
ings, which only occurs on folio 257r 65. The Hagia Sophia
is depicted with two minarets, the wooden minaret on the
west side and the minaret of Mehmed II on the southeast
corner, although the imprecise depiction seems to suggest
the eastern aspect 66. What is striking is the great agreement
with the Düsseldorf reproduction of the Hagia Sophia (east
side with apse and minaret) and the Column of Justinian
with a rider (fig. 9) 67. The same connection between church
(without minarets) and column also appears in the two
other Constantinople depictions in the Liber chronicarum,
but the rider there is incorrectly turned to the left (west). I
have already discussed the question of the possible template
for the Hagia Sophia and Justinian‘s Column, including the
equestrian statue that no longer existed after 1456. It could
not be proved that the Constantinople view by Peronet Lamy
(† 1453), contained in two copies of the Notitia Dignitatum,
goes back to Ciriacus of Ancona, and it could neither be
proven that either of the two copies was known to Schedel
or the illustrators of the Liber chronicarum 68. The Church of
St John the Theologian on Diippion (mistakenly referred to
as Joh<ann>is bapt<is>te) is only illustrated once, but with-
out an inscription, in the above-mentioned (p. 22) Parisian
Buondelmonti vedute from the period after 1457/1458 69.
The stables of Mehmed II, here differentiated as horse and
camel stables (Stabula equitum and Stabula camelorum, re-
spectively), are shown as an open, four-sided courtyard with
gates and windows, as well as on the İstanbul miniature
of Matrakçı Nasuh from 1537 70. In the slightly older Düs-
seldorf depiction it is reproduced as a free-standing build-
ing within a courtyard with two entrances (fig. 2 and 4) 71.
Stephan Gerlach (in İstanbul from 1573 to 1578) described
the stables as follows: »Further I came to the Emperor’s sta-
bles, an exceedingly large, long and wide building, in which
several hundred horses can stand [...] The place where the
horses stand is a building, as otherwise a karavanseray, in
66 Effenberger, Minarette.
67 On the Düsseldorf veduta, only the minaret on the southeast corner of the
Hagia Sophia is indicated, the wooden minaret on the southern staircase tower
of the west side was not visible from the chosen viewpoint.
68 It is the copy of the lost Speyer Codex in Oxford, Bodleian Library, MS. Canon.
Misc. 378 (the Constantinople miniature on fol. 84r), of which another copy
exists in Munich, Bayerische Staatsbibliothek, Clm 794 (the Constantinople
depiction, fol. 167r, see Effenberger, Minarette 197-200, colour fig. 9).
69 Westbrook, Freshfield Folio, wants to identify the building on folio 21 of the
»Freshfield Album« (Cambridge, Trinity College, MS. 0.17.2) with the Church
of St John on the Diippion, but ignores the pictorial sources (and, accordingly,
the literature, which deals with it), which show a double shell octagon with
drum. For an important written testimony, see Grélois, Note. – For the church,
see now Magdalino, The Church of St John the Apostle.
70 İstanbul, University Library, MS. Yıldız Ty 5964, fols 8b and 9a: Yurdaydın, Ma-
trakçı Nasuh pl. 8a.
71 See above n. 60.
the middle a beautiful large open space, a lot of water in
it, and a high house, like a quadrangular tower« 72. The Han
character of the building seems to have been characterised
best on sheet 257r. From the arsenal of Kadırgalimanı only
the roofs of the five ship sheds can be seen behind the sea
wall, but it is confirmed once again that there were only five
sheds. Most striking is the reproduction of the Muchrutas (a
Seljuk pavilion), which apparently still existed within the indi-
cated palace ruins 73. The buildings in Topkapı Sarayı (Domus
mag<ni> turci) and in the serail garden (Viridium) are also
of interest. The missing upper floor of Bab-ı Hümamayun
and the two towers flanking Babüsselam reappear almost
simultaneously on the Düsseldorf view (fig. 2) and on sheet
257r of the Liber chronicarum (fig. 8). The Hagia Eirene (erro-
neously referred to as S. Joh<ann>es Chrysostoma 74) is also
placed only on the Düsseldorf veduta in the first courtyard
of the serail with the inscription S. elini 75. The church in the
corner of the serail appears for the first time on the Düssel-
dorf view in the same place and is labelled here as S. Maria.
I consider it to be the Hodegetria Church 76. The inscription S.
Geor<g>ius must be related to the Georgios Church further
to the north in the Mangana quarter. The damaged church
building (Destruct<i>o antiqua) outside the line, which ac-
cording to the inscription circulus deuast<i>onis marks the
limit of the effects of the storm, must mean the Baruthane
and not the Nea Ekklesia (see n. 60 above) 77. Incidentally,
the view on page 257r illustrates oral reports of the storm,
for the text states: »as the trustworthy Venetians and other
merchants said«. Perhaps the whole representation goes
back to a locally made leaflet that recorded this event 78 and
that the woodcutters have implemented it in their manner,
using the Column of Justinian with the equestrian statue
from an unknown original.
Ottoman Representations
The Kitab-ı Bahriye (»Book of Seafaring«) by the navigator
and cartographer Piri Reis, of which numerous copies have
been preserved, was made between 1521 and 1526 79. The
two-page İstanbul map is from the bird’s-eye view, as are
the Buondelmonti and Vavassore-type views, but as seen
from the north. The illustration includes Constantinople,
Galata-Pera, the Asian side with Üsküdar and Kadıköy and
the Princes’ Islands. The interior of the city is packed with
72 Gerlach, Tagebuch 336B-337A (21 April 1577).
73 Recognised by Asutay-Effenberger, Muchrutas.
74 Berger / Bardill, Representations of Constantinople 20-21.
75 Effenberger, Illustrationen 23-24 no. 2 fig. 32.
76 Effenberger, Illustrationen 27-28 no. 9 fig. 32. – Grotowski, The Hodegon.
77 Alternatively: Berger / Bardill, Representations of Constantinople 23.
78 Berger / Bardill, Representations of Constantinople 15: »apparently based on a
sketch made on the spot by Schedel’s informers«.
79 Orbay, Istanbul viewed 117-289. – Soucek, Piri Reis 132-135. – Effenberger,
Illustrationen fig. 11 (copy in Berlin, Staatsbibliothek, Orientabteilung, Sign.
Dietz A fol. 57, c. 1663-1724/1725).
Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
Fig. 10  Detail from the depiction of Constantinople in the Kitab-ı Bahriye of Piri
Reis (London), showing the Harbour of Theodosius with free-standing tower. –
(London, Khalili Collection. From Asutay-Effenberger, Landmauer fig. 47).
mosques and houses in a condensed manner, with the most
important buildings and city gates labelled with the Ottoman
names 80. The Harbour of Theodosius is reproduced as on
the Buondelmonti copies in the state before the complete
closure of the alluvial bay. It was only depicted with the
free-standing tower in the sea, which was also erroneously
represented as a round tower (fig. 10) 81. This state of the
harbour, which was long obsolete at the time of origin of the
map, is therefore likely to date back to an unknown original,
which must be older than the prototype of the Vavassore
view of 1479/1481 82. A copy of the Buondelmonti type can
probably be ruled out, especially since neither moles nor signs
of sedimentation can be identified.
The İstanbul miniature by Matrakçı Nasuh from 1537 fol-
lows a different depiction principle 83. İstanbul is seen from
the side of the Theodosian walls, thus reproduced from the
west 84. Galata-Pera is only to be considered correctly if the
double leaf (each 31.2 cm × 22.5 cm) is turned by 90° in a
clockwise direction 85. Many buildings can be identified and
prove the high value of this veduta as a witness for the state
of the city at the time of Suleiman the Magnificent 86. Within
the two cities, the individual buildings are usually arranged
horizontally, but the painter had to solve the problem of
80 Asutay-Effenberger, Landmauer 216-223 fig. 47 (London, Khalili Collection). –
Asutay-Effenberger, Kitâb-ı Bahriye.
81 In the London copy, the tower has the inscription »Kulle-i Hamza«, see Soujek,
Piri Reis 134. – Asutay-Effenberger, Landmauer 219.
82 Effenberger, Illustrationen 32.
83 Orbay, Istanbul viewed 47-67. – Halbout du Tanney, Istanbul.
84 Yurdaydın, Matrakçı Nasuh pl. 8b.
85 Yurdaydın, Matrakçı Nasuh pl. 8a.
86 Denny, Plan of Istanbul (often with wrong identifications).
29

how to depict walls and alignments of buildings running in
a west-east direction. Thus, the towers and curtain walls of
the sea walls at the Golden Horn along the shore are shown
standing upright. On the Sea of Marmara, the painter tried
to make a perspective view, which reached to the sphendone
of the Hippodrome, but on the vertical coastline, the towers
seem to »topple over« into the sea. In the case of the walled
and almost square-shaped district of Langa Bostanı, the eight
30 Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
Fig. 11  Detail from the view of
Istanbul by Matrakçı Nasuh, show-
ing Kadırgalimanı and Langa Bos-
tanı. – (From Halbout du Tanney,
Istanbul fig. 21).
towers stand on three sides on an imaginary level and tilt
inwards, while the two large corner towers on the sea side
»topple over« back into the sea (fig. 11). There are lovely
details inside, such as a large bed of flowers and vegetables,
three trees and two draw wells. The inner basin, the retaining
wall and the large gate of Kadırgalimanı are depicted, but the
artist made a mistake by moving the five ship sheds to the
opposite end of the inner bay.
 The Hünername (»Book of Skills«) created by Seyyid Lok-
man in 1584/1585 for the Sultan Murad III (1574-1595) con-
tains a map of İstanbul (44.0 cm × 27.5 cm) attributed to the
painter Veli Can 87. In addition to the actual city, the map
extends from the Golden Horn to the confluence of the
combined rivers of Alibey suyu and Kağıthane suyu, a large
part of Galata with the Tersane-i Amire of Kasımpaşa, the
villages Hasköy and Sütlüçe, the cult district of Eyüp, the
settlements on the foreland of the Theodosian walls and a
corner of Üsküdar. In the view, a different kind of representa-
tion of the building is encountered, which actually forces the
image to be rotated and viewed from several angles. This is
exactly how the draughtsman must have done his work and
chosen different city points 88. The lower third of İstanbul is
viewed from the Marmara side, the upper part assumes the
view from the Golden Horn. The fact that the buildings are
Summary / Zusammenfassung
Constantinople / İstanbul: The Early Pictorial Sources
The illustrations included in the individual chapters on the
harbours have a varying degree of testimonial value, both
in terms of the state of the city of Constantinople / İstanbul
and also terms of the representation of the harbours and
landing places. This applies especially to the early city views
from the fifteenth and sixteenth centuries. In this chapter, the
harbours on the vedute of the Buondelmonti and Vavassore
types, of Hartmann Schedel, as well as several early Ottoman
views (Matraçı Nasuh, Piri Reis, Lokman’s Hünername) will
be discussed. For the illustrations of the Buondelmonti and
Vavassore types, possible underlying pictorial sources will be
analysed. The Düsseldorf view of İstanbul in Cristoforo Buon-
delmonti’s Liber insularum archipelagi of 1485/1490, which
already conveys the state of the city at the time of Mehmed II
or Bayezıd II, proves to be significant in many respects.
87 İstanbul, Topkapı Sarayı Kütüphanesi, Hazine 1523, vol. 1, fols. 158b. 158ar. –
Anafarta, Hünername pl 37. – Orbay, Istanbul Viewed 73-116. – Effenberger,
Illustrationen fig. 12.
Constantinople / İstanbul: The Early Pictorial Sources  |  Arne Effenberger
turned upside down in each case may not have disturbed the
observers, as the city is actually made to be experienced by its
two main sides. Nevertheless, the view is relatively reliable in
topographical terms. Although labels are missing, most build-
ings can be identified. The apparent doubling of the walled
Harbour of Theodosius could be clarified 89. The reproduction
of Kadırgalimanı is inaccurate 90.
In summary, only the Düsseldorf view (fig. 2) reflects the
contemporary situation of both the former Harbour of Theo-
dosius and the situation of Kadırgalimanı during the reign of
Mehmed II in a way that corroborates or complements the
written sources. This is due to the fact that with this veduta
we have a contemporary pictorial document that was made
by a draughtsman who was well-versed in local circumstances
(if not particularly gifted) and whose intention was no longer
to illustrate the Constantinople chapter of the LIA.
Konstantinopel / İstanbul: die frühen bildlichen  
Zeugnisse
Die den einzelnen Hafen-Kapiteln beigegebenen Abbildun-
gen haben einen unterschiedlichen Zeugniswert sowohl in
Bezug auf den Zustand der Stadt Konstantinopel / İstanbul
als auch mit Blick auf die Darstellung der Häfen und Anle-
gestellen. Dies betrifft vor allem die frühen Stadtansichten
aus dem 15. und 16. Jahrhundert. In dem Beitrag werden
die Hafenanlagen auf den Veduten vom Buondelmonti- und
vom Vavassore-Typus, Hartmann Schedel sowie einige frühe
osmanische Ansichten (Matraçı Nasuh, Piri Reis, Hünername
des Lokman) untersucht. Für die Darstellungen vom Buon-
delmonti- und Vavassore-Typus werden die möglicherweise
zugrundeliegenden Bildquellen behandelt. Als in vielerlei Hin-
sicht aussagekräftig erweist sich die Düsseldorfer İstanbul-
Ansicht im Liber insularum archipelagi des Cristoforo Buon-
delmonti von 1485/1490, die bereits den Zustand der Stadt
zurzeit Mehmets II. bzw. Beyazıts II. überlief.
88 Orbay, Istanbul Viewed 90-114.
89 Effenberger, Illustrationen 32.
90 See Heher, Harbour of Julian fig. 13, in this volume.
31
Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
Some Reflections on the Archaeology
of the Late Antique and Byzantine Harbours
of Constantinople
This chapter aims to complement the other contributions of
this volume 1 with particular remarks on the physical remains
of the harbours of Constantinople. Due to the impossibility of
covering the entire spectrum of archaeology within a chapter,
the following archaeological commentary will focus primarily
on the architecture of harbour facilities per se, which consist
of features such as quay structures, breakwaters with mole
or wall superstructures, as well as jetties and pier remains. In
striking contrast to the relative abundance of historical ac-
counts, the visual and archaeological knowledge of harbour
structures along the coasts of the Golden Horn and the Sea of
Marmara was for a long time limited to sparse visual evidence,
such as photographic illustrations, maps, plans and historical
depictions, together with short references and rather vague
field notes 2. In fact, it was not until 2004 that a comprehen-
sive insight into the harbours of Byzantine Constantinople,
notably the harbour of Theodosius at Yenikapı, Chrysopolis at
Üsküdar and Neorion / Prosphorion at Sirkeci was possible for
the very first time due to the large-scale rescue excavations
conducted prior to the so-called Marmaray-Metro Construc-
tion Project 3. Following the geographical order of the Volume,
the paper is structured on the basis of the archaeological data.
Hence, starting with the site of Yenikapı.
The Theodosian Harbour at Yenikapı
With an exceptionally large excavation area of 58 000 m²
(fig. 1), the archaeological site at Yenikapı revealed, not only
a total number of 37 shipwrecks of the Early to Late Byzantine
periods 4, but also a multitude of architectural remains. The
latter are situated both at the eastern and the very western
1 A first German version without the present article was published in 2016: Daim,
Häfen.
2 Dark, Harbours 152-154; Demangel, Contribution 42. 46-47; Krischen, Land-
mauer tab. 18; Schneider / Meyer-Plath, Landmauer tab. 25a-b; Mamboury / Wie-
gand, Kaiserpaläste, tab. XXVIII-XXX, XXXV-XXXVI; Mango, Spolia figs 1-3;
Müller-Wiener, Häfen, tab. 48,2; Simeonov, Brachialion, in this volume, figs 3-4.
12; G. Simeonov, Hebdomon figs 1. 14, in this volume.
3 Kızıltan, Yenikapı, Sirkeci and Üsküdar.
4 For a detailed study of the shipwrecks see Kocabaş, Yenikapı Shipwrecks; Ko-
cabaş et al., Collection; Pulak / Ingram / Jones, Byzantine Shipwrecks; Pulak et al.,
Shipwrecks of Yenikapı. A short but excellent overview with an historical analysis
is provided by Külzer, Harbour of Theodosius esp. 84-89, in this volume.
In: Falko Daim  ·  Ewald Kislinger (eds), The Byzantine Harbours of Constantinople. Byzanz zwischen Orient und Okzident 24 (Mainz 2021).
DOI: https://doi.org/10.11588/propylaeum.911.c12065 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
end of the harbour excavation area. Amongst the various
harbour structures brought to light, the most striking features
form two massive jetties located in the eastern harbour basin
(fig. 2) 5. Both jetties are oriented in a north-south direction
corresponding to a perpendicular position to the northern
shoreline of the late antique harbour basin (fig. 3). In con-
trast to the very poor state of preservation of the western
jetty, the eastern one is in surprisingly good condition and
completely preserved over a length of 35 m and a total width
of 4 m (fig. 4) 6.
The structure consists of two different parts: a solid and
homogeneous foundation, which has a uniform and linear
shape; and a superstructure of large ashlar blocks (fig. 5).
The foundation is characterised by a compact composition
of mortar mixed with rubble stones and ceramics (figs 6-7) 7,
resembling the construction method of opus caementicium.
Considering the unique conditions in the marine environ-
ment, it can be assumed that the construction of the foun-
dation required a certain type of hydraulic concrete, mixing
quicklime, seawater and an aggregate as a mortar-binding
material 8. Whether the aggregate used for the concrete
composition consists of pozzolanic mortar, the so-called pu-
teolanus pulvis (a volcanic sand from the Gulf of Naples near
Puteoli) 9, or any other volcanic ash or aggregate, still needs
to be investigated through archaeometric analysis. Although
Brandon aptly suggests that the concrete foundation of the
eastern jetty represents a structure that can no longer be
defined as »Roman marine concrete« 10, it nevertheless fol-
lows Roman harbour construction techniques. As described
by the Roman architect and engineer Vitruvius Pollio, as well
as later by the Byzantine scholar and historian Procopius of
Caesarea, hydraulic concrete installations protruding into the
  5 A jetty forms a permanent solid structure built out into the sea or harbour basin
as part of a harbour or dockyard alongside which ships could berth for loading
and unloading activities: Ginalis, Byzantine Ports 35; Dear / Kemp, Ships and the
Sea 290-291.
  6 Ercan, Yenikapı 121.
  7 Gökçay, Architectural Finds 177.
  8 For the technology and character of Roman marine and hydraulic concrete see
Blezard, Cements; Brandon et al., Building for Eternity 1-4. 141-187.
  9 As has been used for example at Roman Imperial harbours such as Caesarea
Maritima, Pompeiopolis, Cosa, etc.: Brandon et al., Building for Eternity 73-81.
94-101; McCann, Cosa; Oleson, Technology; Raban, Caesarea Maritima 64 ff.
10 Brandon et al., Building for Eternity 136.
33
water, such as jetties or moles, were constructed by using rec-
tangular wooden formworks or chests (as Procopius refers to
them) 11. Such wooden formworks or caissons were prepared
on land and subsequently sunk into the water in order to be
placed on the seabed for the filling of the hydraulic concrete
mixture 12.
Remains of such wooden caissons have actually been
preserved at the eastern jetty at Yenikapı, indicating that the
feature is composed of a series of individual concrete masses
(figs 4. 6) 13. Four samples from parallel vertical boards of the
wooden formwork had been dated through dendrochron-
ological analysis to the period between AD 657 and 786 14.
Hence, the construction of the jetty may well be allocated to
further historically documented harbour works such as the
construction or repair of the harbour fortification surrounding
the harbour basin at the turn of the seventh to the eighth
century or further restoration measures in the first half of the
ninth century 15.
A dating to the Middle Byzantine period is further sup-
ported by the upper construction part, which sits on the
homogeneous concrete foundation. This superstructure con-
sists of large ashlar blocks (fig. 8). These, however, do not
represent only uniform building materials, but also a mixture
of various re-used blocks fitted in for the construction of
the walking level. Accordingly, apart from mostly limestone
blocks of different shapes, marble blocks and even three
spolia blocks are used. The latter pertains to the fragments
of a frieze block decorated with a band of acanthus leaves
confined by strips of egg-and-dart and Lesbian cymatium
(fig. 9) 16. Based on the decorative style, a terminus post
quem of the mid-fifth to the mid-sixth century may be con-
sidered for the three decoration fragments 17. As such, the
superstructure again indicates that the construction of the
eastern jetty may not be dated earlier than the beginning
of the seventh century. Additionally, the superstructure does
not form a continuous level of ashlar blocks. Instead, the
blocks were merely placed at the edges of each concrete unit,
thus forming chambers. The chambers were subsequently
filled again with a rough conglomerate of quarry stones and
11 Vitruvius, De Architectura V. 12. 3 (129 Rose / Müller-Strübing); Prokopios, De
Aedificiis I 11, 18-20 (IV 44 Haury / Wirth).
12 Brandon et al., Building for Eternity 189-222.
13 Ercan, Yenikapı 122-123; Gökçay, Architectural finds 177; the up-to-five
preserved units show an inclination towards the harbour basin of +1.15 m,
+1.21 m, +1.15 m, +1.42 m and +1.57 m: Ercan, Yenikapı 123.
14 Kuniholm et al., Of Harbors and Trees 63.
15 Külzer, Harbour of Theodosius 40, in this volume; Müller-Wiener, Häfen 9.
16 Ercan, Yenikapı 121.
17 Comparative examples from the sea walls, as well as the sea gate at the Bou-
koleon Palace, suggest a date during the reign of Emperor Justinian I (6th c.):
Mamboury / Wiegand, Kaiserpaläste tab. XVII-XVIII; Mango, Spolia 648 fig. 7.
18 Ginalis, Anthedon.
19 Ibid.
20 Ginalis, Anthedon; Ginalis, Byzantine Ports 190; Knoblauch, Ägina 73; Paris,
Lechaion 10-11; Rothaus, Lechaion 295-296; Schäfer, Larymna 533-537;
Schläger / Blackman / Schäfer, Anthedon 36, Abb. 14; Triantafillidis / Kout-
soumba, Aegina 169.
21 Ginalis, Anthedon; Ginalis, Byzantine Ports 176-177. 193. 238-239. 244-245;
Karagiorgou, Urbanism 31. 168 ff.; Trombley, Boeotia 991-992.
34 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
mortar, most likely another hydraulic concrete composition
(figs 10-11).
It is very likely that the chamber system technique was
intentionally chosen for achieving a robust construction, but
pertaining an inexpensive technique with available construc-
tion materials and labour means 18. Accordingly, this may well
correspond to a time when the Byzantine Empire was in need
of swift action, as it was facing serious economic difficulties 19.
Yet, when did this occur?
The technique used in Yenikapı has counterparts in a num-
ber of harbour sites primarily along the central Greek coasts,
such as the harbours of Anthedon, Larymna, Theologos or
Aegina and the outer harbours of Thessalian Thebes, but
also at the Corinthian harbour of Lechaion 20. The marked
proliferation of these harbours seems to be directly linked
to the growing importance of Boeotia and Thessaly as major
producers and suppliers of grain and likewise other agricul-
tural products, particularly from the seventh century onwards,
in relation to the well-known consequences of the Arab con-
quests 21. The increasing role of this new maritime network
has clear reflections in Constantinople, and particularly in its
largest harbour located in Yenikapı.
A reference to a warehouse / granary called Horrea or Horion
Lamias situated on the eastern side of the Theodosian har-
bour from the seventh century onwards is particularly note-
worthy within the historical context 22. This granary 23, identi-
fied with the so-called Horrea Alexandrina listed in the Notitia
urbis Constantinopolitanae from the fifth century AD 24, not
only indicates continuous trade and shipping activities in the
Theodosian harbour up to the late Middle Byzantine period,
but also its possible close relationship to the harbour network
of Central Greece 25. This phenomenon possibly evinced by
the etymology of granary’s name, Lamia, which has been
previously explained by a female monster 26. Nevertheless, as
the Horrea Alexandrina signified the shipment of grain from
Alexandria in Egypt, it is very likely that the Horion Lamias
is associated with the city of Lamia 27  – thus indicating the
shipment of grain from central Greece, as a substitute of
22 Miracula Artemii (Crisafulli / Nesbitt) 107 (16); Patria Konstantinoupoleos 51. 85
(II 179, 246 Preger).
23 For warehouses, granaries and other commercial facilities see Ginalis, Byzantine
Ports 48-54.
24 Notitia urbis Constantinopolitane X 6. 9 (237 Seeck); Magdalino, Constantino-
ple 23; Mundell Mango, Commercial Map 200-201 fig. 4; Kislinger, Better and
Worse Sites 9-10, in this volume.
25 Magdalino, Grain Supply 37.
26 Ercan, Yenikapı 78; Janin, Constantinople 351-352.
27 Until the middle of the 6th c. the important Phthiotian city, which during the
Byzantine era belonged to the province of Thessaly, was known with its ancient
toponym »Lamia«; the bishopric was refounded in the 8th-9th c.: Koder / Hild,
Hellas und Thessalia 53-54. 81. 283-284. Written accounts adopted the Slavic
origin toponym »Zetounion« (from the Palæoslavic word »zito«, meaning
»grain« or »cereal crop«) only after the 9th c. However, its wider surrounding
agriculturally fertile area remained known and associated with the toponym
»Lamia«: Avramea, Thessalia 199; de Rosen, Rhomanian Boeotia 138-139;
Karagiorgou, Urbanism 94-95. 107-110. 113; Pallis, Lamia 59.
Egypt. In this respect, the creation of the theme of Hellas in
AD 695 28 could again serve as an historical reference point for
the construction of the eastern jetty at the turn of the seventh
century to the eighth century 29.
Some of the ashlar blocks feature small notches of either
rectangular or trapezoidal shape, also known as dovetail
grooves (fig. 12). Metal clamps, which were formed by pour-
ing liquid lead into these grooves, provided a strong bonding
between the individual blocks in order to achieve a high
stability and long-lasting resistance 30. However, the isolated
appearance of grooved ashlar blocks indicates that the use of
metal clamps was not an essential measure for the stability of
the jetty’s upper part. Consequently, certain blocks seem to
have been removed from their original place in the structure,
rendering the notches more or less unnecessary. In fact, the
use of dovetailed lead fixings seems to be a characteristic
of ancient harbour architecture, for instance known from
Caesarea Maritima 31. Possibly taken from an earlier harbour
facility, it presumably originated either from another jetty
or an older quay installation along the eastern end of the
harbour (see below).
Grooved stone blocks in a similarly re-used context can
also be found at other Byzantine harbour sites, such as at
Thessalian Thebes, Lechaion or along the Küçükçekmece
Lake 32. Similar to Anthedon, the chambers of rubble stones
and mortar were finally covered with a last layer of lime-
stone ashlar blocks and stone slabs, respectively (fig. 13) 33.
Whether the jetty bore a further superstructure of brickwork
remains unknown. During the Byzantine era, the Roman
tradition of combining concrete with brick for the construc-
tion of arched harbour structures most likely continued (as
shown for example by Cristoforo Buondelmonti’s depiction
of Constantinople in his Liber insularum archipelagi) 34. Since
the brickwork usually rests directly on the concrete founda-
tion, in this case, such a structure on top of the ashlar block
chambers should not be excluded.
Finally, in contrast to the usually rectangular-shaped
wooden chests, a pentagonal shape was chosen for the
southernmost caisson. Visually, this results in giving the jetty
a pointed end (fig. 4). Due to this singular ground plan, it has
been suggested that the last part of the jetty may be associ-
ated with the previous existence of a lighthouse 35. However,
28 Koder / Hild, Hellas und Thessalia 57; Živcović, Date of the Creation 142-143.
About the considerable extension of the Slavic controlled territories in central
and southern Greece from the late sixth to the early 9th c., see Koder, Sied-
lungsgebiete; Kislinger, Regionalgeschichte als Quellenproblem and Kislinger,
Dyrrhachion.
29 Ginalis, Anthedon.
30 Ercan, Yenikapı 121; Ginalis, Byzantine Ports 184. 242.
31 Kingsley, Barbarian Seas 136; Raban, Sebastos, Royal Harbour 115. A connec-
tion between the lead clamps and the so-called ἄργυρος χυτός (»liquid silver«
or rather quicksilver), referred by the Late Byzantine historian Pachymeres is
highly questionable and has been rightly doubted by Heher, Harbour of Julian
60, in this volume.
32 Aydıngün, Excavation Site 17 figs 49-50; Aydıngün / Aydıngün / Öniz, Küçükçek-
mece 440-441; Ginalis, Byzantine Ports 184, Ill. II.II.48b.
33 Ginalis, Anthedon; Schläger / Blackman / Schäfer, Anthedon 47.
Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
beyond the fact that no evidence of such a structure could be
determined whatsoever 36, a lighthouse or lid beacon within
the interior harbour zone is not plausible considering its lo-
cation within the harbour basin. As such, it can be suggested
that the pentagonal shape, together with the incorporated
ornamental spolia, rather served purely visual aesthetics 37.
A further wall joins the eastern jetty at its southern end,
which extends the structure approximately 20 to 25 m to-
wards the south (fig. 14). The wall consists of one row with
two preserved layers of large, re-used ashlar blocks and rub-
ble that are set in a system of headers (fig. 15). A second row
most probably existed. This implies a maximum total width of
around 2 m (half the size of the above-described Middle Byz-
antine jetty). Without any use of mortar binding, the blocks
seem to have been placed loosely on the ground, providing
the impression of a rather provisional construction. On top of
the well-worked ashlar blocks, undressed stones are placed
up to the height of the walking level of the adjacent eastern
jetty. Despite the big difference in size to the latter, the struc-
ture should also be identified as a jetty.
However, based on the rough construction technique, as
well as the fact that it forms an annexe to the Middle Byzan-
tine equivalent, a much later date has to be assumed. This is
supported by the processing of re-used ashlar blocks, which
were most likely dismantled from the eastern jetty to its north.
The extension of the eastern jetty towards south is due to the
constant siltation process of the harbour basin by the Lycus
river (Bayrampaşa Deresi today), which emptied into the bay
of the Theodosian harbour (fig. 3) 38. The loose arrangement
of ashlar blocks, without any mortar binding, indeed indicates
that they must have been placed either on dry ground or
in relatively shallow water. As such, the water depth in the
harbour basin must have dropped by a large extent and the
navigable sea level must have retreated farther south at some
point after the ninth century. This obviously required building
measures to reach the necessary draught for the docking of
the vessels. Based on the wreck finds in that area, the eastern
part of the eastern harbour basin remained in use for harbour
activities at least until the eleventh century 39.
In contrast to the general assumption that the Theodosian
harbour must have been entirely silted up by the end of the
twelfth century at the latest, Külzer rightly suggests a limited
34 Aidoni et al., Journeys 22; Effenberger, Pictorial Sources fig. 1, in this volume.
For arched harbour structures in Roman times see Blackman, Ancient harbours
II 197. 202 ff; Blackman, Sea Transport 648-649.
35 Kocabaş, Theodosian Harbour 25.
36 Ercan, Yenikapı 134-135.
37 Ibid. 121.
38 Ercan, Yenikapı 86 fig. III. 13; Külzer, Harbour of Theodosius 41, in this vol-
ume; Pulak / Ingram / Jones, Byzantine Shipwrecks 39. The existence and strong
impact of the Lycus River on the development of the Theodosian Harbour is
also reflected in medieval drawings of Constantinople, such as in the Liber in-
sularum archipelagi by Cristoforo Buondelmonti: Effenberger, Pictorial Sources
figs 1-2, in this volume.
39 Kocabaş, Yenikapı Shipwrecks 31 fig. 5; Külzer, Harbour of Theodosius 47-48;
Pulak / Ingram / Jones, Byzantine Shipwrecks fig. 3.
35
but continuous use far into the Late Byzantine period 40. This
may be supported, not only by written accounts, but also by
the existence of Late Byzantine kilns, supposedly unearthed
in close vicinity to the eastern jetty. Located just northeast of
the jetty, the kilns may be related to this very last phase of at
least minor harbour activities up to the fifteenth century. A
further indication for a continuous use even far beyond that
is given by the latest archaeological activities of the Istanbul
Archaeological Museum in the eastern part of the Yalı Mahall-
esi area (at the south-eastern end of the Theodosian harbour
area) 41. During the archaeological excavations, an approxi-
mately 40 m long and 4-5 m wide jetty was uncovered, be-
longing to the very last construction phase at the Theodosian
harbour during the Ottoman period 42. Interestingly enough,
the jetty again features a construction system using wooden
chests filled with a conglomerate of mortar mixed with rubble
stones, which resembles a certain hydraulic concrete mixture.
Finally, despite its at least partial rededication around the
twelfth century, similar to the site of Chrysopolis, the con-
tinuous use of the harbour area far into the Late Byzantine
period and even beyond can be seen here as well (see further
down). This is verified by harbour construction works dating
as late as the Ottoman period. A 1.95 m wide jetty (Kibotos
Iskele) inclining towards the sea can be allocated to these late
harbour works (fig. 16).
The second prominent jetty is located in the western part
of the eastern harbour basin (fig. 2). The roughly 20 m long
structure is in very poor condition compared to its eastern
counterpart despite its massive appearance. Nevertheless, the
archaeological remains allow us to reach to firm conclusions
regarding its architecture, its structural composition and thus
its chronology. The jetty consists of three preserved solid and
homogeneous masses (fig. 17). These show a compact con-
glomerate of mortar, mixed with rubble stones and ceramics,
which again resembles a hydraulic concrete mixture. Whether
that comprises Roman pozzolanic mortar or any other vol-
canic ash or other aggregate, has yet to be examined here,
too. Despite the apparent visual resemblance to the eastern
jetty, the structural composition of the concrete bears some
differences. While the concrete mixture of the eastern jetty
shows a high percentage of small to middle-sized rubble
stones and a comparatively low percentage of mortar (fig. 7),
the composition of the western jetty reveals a much higher
percentage of mortar into which middle-sized to large boul-
ders were embedded (fig. 18). On a closer examination, one
40 See Külzer, Harbour of Theodosius, in this volume; Ercan, Yenikapı 62. 92. 96.
118; Kocabaş, Theodosian Harbour 32; Magdalino, Maritime Neighbourhoods
215.
41 For further information on the archaeological fieldwork conducted in this area,
see Öncü / Çölmekçi, Istanbul Boğazı; Öncü / Çölmekçi, Istanbul Boğazı 2016.
42 Akkemik et al., Dendroprovenancing.
43 Ercan, Yenikapı 59 fig. III.2; Külzer, Harbour of Theodosius fig. 4.
44 YK 37 forms the northernmost wreck find and belongs to the earliest group of
vessels retrieved from the Theodosian Harbour: Kocabaş, Yenikapı Shipwrecks
34 fig. 5.
45 Different to permanent solid jetties, the pier forms a structure of timber sup-
ported on wooden piles. Piers were constructed in addition to jetties in order to
36 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
can observe that the embedded stones are not waste quarry
stones but whole river stones. These most probably derive
from the nearby Lycus River, which emptied into the eastern
harbour basin 43.
The use of construction material from the immediate vi-
cinity may point to the earliest construction phase and the
foundation of the harbour in the Early Byzantine period. This
is supported, not only by the shipwreck YK 37 (dated to
the fifth century) in its immediate vicinity (figs 19-20) 44, but
also by the implementation of Roman engineering and con-
struction techniques using rectangular wooden formworks.
Although no physical remains of caissons were discovered
at the western jetty, the gaps between the individual masses
again clearly indicate the utilisation of such chests. In addition,
the colossal dimensions and compactness of the hydraulic
concrete masses (compared to the eastern jetty) imply a much
earlier construction date on their own.
Unfortunately, apart from large stone blocks scattered
around the concrete masses, as well as layers of massive
ashlar blocks at the southern front of the jetty, no further
construction components have been preserved that could
provide any additional indications (fig. 21). As for the south-
ern end of the jetty, it should be noted that the ashlar blocks
do not rest continuously on the hydraulic concrete mass (as
is the case at the eastern jetty). Instead, they give the im-
pression that they are fitted into the washed-out and eroded
concrete. This building measure could have aimed for two
possible purposes: either to stabilize the jetty against the
risk of collapsing; or, more likely, to extend the structure
further south. As such, the massive ashlar blocks may be
considered as later additions. A step-like arrangement of the
ashlar blocks towards a wooden pier (see below) supports this
interpretation. Marble column pieces and a marble impost
block with the monogram of Emperor Justinian I have been
unearthed immediately in front of the jetty (fig. 22), pro-
viding a terminus ante quem of the mid-sixth century for its
erection. Consequently, it may be assumed that the western
jetty was erected as early as the end of the fourth century or
the beginning of the fifth century and underwent repair or
extension during the sixth century. This is further supported
by another type of harbour infrastructure: wooden piers 45.
Throughout the harbour basin, a large number of wooden
piles belonging to piers have been brought to light, ranging
from the fifth to the fifteenth centuries based on dendro-
chronological analysis 46. A great majority of them is oriented
provide additional mooring space for ships within the harbour basin. As shown
for example by the late antique to medieval harbour of Olbia, piers were used
equally to jetties within harbour areas, leading into the basin at a right angle
to the shoreline: Dear / Kemp, Ships and the Sea 427; Ginalis, Byzantine Ports
35-37; Kingsley, Barbarian Seas 89-90.
46 Kuniholm et al., Of Harbors and Trees 47; it has to be mentioned that the
dating of the wooden remains always refer to the time of their cutting and not
necessarily to their immediate use for construction. A certain time period has
obviously to be calculated from the time of cutting the woods, the transport
and processing of the material to their use for building activities.
in a north-south direction (fig. 23). In line with the distribu-
tion of shipwrecks 47, the earliest pier constructions have been
discovered at the northern end of the harbour basin, pointing
to their connection to the harbour’s quay area. According
to the analysis of a series of posts, almost all wooden piers
show multiple phases, which correlate with numerous repairs
as well as enlargements (thus reaching total lengths of up to
over 40 m) going along with the siltation process towards
south and southeast 48. The longest-lasting pier with a usage
of over 80 years and three phases of repair / extension (four
phases in total) constitutes the so-called »Marmaray İskele 1«
at the western end of the harbour 49. While the earliest phase
dates to around the year AD 527, its latest posts are from
around the year AD 610. A date around AD 553 is also given
for the wooden pier connected to the southern end of the
western jetty (fig. 24) 50. This again provides a terminus ante
quem of the mid-sixth century for the erection of the western
jetty. Simultaneously, dendrochronological analysis attests to
its continuous use up to the ninth century, which is similarly
confirmed by the nearby shipwrecks YK 27, YK 28 and YK 32,
dated to the seventh to ninth centuries (figs 19-20) 51.
At the western end of the harbour basin a further se-
ries of harbour installations has been uncovered (figs 3. 25).
The wealth of different overlapping facilities provides a very
complex picture, leading scholars to different interpretations
and still puzzling the excavators. Concerning harbour-related
structures, the most striking feature forms a quayside 52. Due
to the limitation of the excavation area, only a total length
of 25.50 m could be revealed (fig. 26). The 2.80 m wide
quay shows a southwest-northeast orientation and consists
mostly of a single row with 1-2 layers of ashlar blocks 53.
The latter, however, are not comprised of standardised or
uniform construction material, but rather randomly ar-
ranged, 2.75 m × 1.35 m large stones. Interestingly enough,
these stones represent almost exclusively re-used material
of bossage and local dressed stone slabs (fig. 27). The com-
pilation of re-used stone material is supported by the use of
two inscribed spolia blocks (figs 28-29a) 54. Similar to the
extension of the eastern jetty (see above), again no mortar
binding was used, and the blocks are only loosely placed on
the ground and on top of each other 55.
47 Kocabaş, Yenikapı Shipwrecks fig. 5.
48 Gökçay, Architectural Finds 168. 176; Kuniholm et al., Of Harbors and Trees
58-63. 66-77; Pearson et al., Dendroarchaeology 3407 fig. 8.
49 Gökçay, Architectural Finds 168; Kuniholm et al., Of Harbors and Trees 67;
Pearson et al., Dendroarchaeology 3408.
50 Pearson et al., Dendroarchaeology tab. 1.
51 Kocabaş, Yenikapı Shipwrecks 21. 23 fig. 5; Kuniholm et al., Of Harbors and
Trees 61; Pearson et al., Dendroarchaeology; Külzer, Harbour of Theodosius 46,
in this volume.
52 A quay forms a projection along the shoreline of the harbour, usually con-
structed of stone masonry. The solid structure constitutes the main facility for
the accommodation of ships to load and unload cargo or embark and disem-
bark passengers: Dear / Kemp, Ships and the Sea 450; Ginalis, Byzantine Ports
32-34.
53 Ercan, Yenikapı 120; Gökçay, Architectural Finds 170; Kocabaş, Theodosian
Harbour 25.
54 Ercan, Yenikapı 120; Gökçay, Architectural Finds 170.
Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
One of the ashlar blocks features a 10 cm wide hole
pierced horizontally through the stone, whereas one of
the inscribed spolia blocks points to a second perforation
(fig. 29a-b). Rather than interpreting them as being part of
a lifting device 56, the holes were intended for the mooring
of ships. Besides vertically projecting bollards, perforated
stone blocks or so-called »mooring stones« formed the most
commonly used device for berthing ships since classical an-
tiquity 57. By piercing the blocks of the quay’s frontal façade,
the mooring device was incorporated into the wall as a single
architectural unit with the quay.
As for the dating of the quay line, it has been suggested
that the harbour facilities at the western end belong to the
earliest construction phase, possibly dating to the initial
building project of the Theodosian harbour between AD
390 and 425 58. However, one gets the impression that the
rough and seemingly provisional construction does not reflect
a representative installation for an imperial harbour of the
fourth / fifth century – especially given the fact that compara-
bly large harbour sites, such as Caesarea Maritima, Demetrias,
Thessalian Thebes, Corinth’s eastern harbour of Lechaion or
Ephesus, show a far more elaborate architecture 59. Prima
facie, an earlier construction period significantly preceding
the Byzantine era may at first be suggested by the building
material and the inscribed spolia used 60. On closer examina-
tion, however, the construction assembly is of clearly re-used
context, which suggests a rather later date. This is in fact
supported by further excavation works undertaken in the
Light Rail System area north-west of the Yenikapı site.
Among a series of building remains, which are roughly
dated between the sixth / seventh and the ninth century, a
13 m long and 3 m wide structure has been unearthed 61.
Running in an east-west direction parallel to the Theodosian
harbour and extended by wooden piles suggests an identifi-
cation as a jetty with a pier projection belonging to a further
harbour infrastructure. Whether its preserved part is sitting
on a hydraulic concrete foundation within wooden caissons
is not known to the authors. The partly destroyed jetty shows
a solid architecture consisting of ashlar blocks with mortar
binding and a flat surface covered by a thick concrete layer
(fig. 29c). Hence, it is more reminiscent of the elaborate and
55 Ibid.; Kocabaş, Theodosian Harbour 25.
56 Kocabaş, Theodosian Harbour 25.
57 For mooring devices see Blackman, Bollards 115-122; Ginalis, Byzantine Ports
38-43.
58 Gökçay, Architectural Finds 170; Külzer, Harbour of Theodosius 39, in this vol-
ume.
59 Ginalis, Byzantine Ports 162-193; Kingsley, Barbarian Seas 132-138; Külzer,
Ephesos 49-57; Paris, Lechaion; Rothaus, Lechaion; Steskal, Ephesos; https://
www.theguardian.com/science/2017/dec/14/new-underwater-discover-
ies-in-greece-reveal-ancient-roman-engineering (accessed 8 July 2019).
60 The incorporated inscribed spolia blocks provide a terminus post quem of the
Roman Republican period for the erection of the quayside. While Gökçay dates
the incorporated inscribed spolia blocks to the 4th c. BC, Ercan suggests a date
between the third and the second centuries BC: Ercan, Yenikapı 120; Gökçay,
Architectural Finds 170.
61 Kızıltan, İstanbul Kazıları 362.
37
representative architecture of imperial harbour installations
than the quay facility described above. Its location north of
the maritime circuit wall enclosing the Theodosian harbour
basin (see further down) implicates the existence of a har-
bour site prior to the foundation of the Theodosian harbour.
As such, a date to the fourth century or even earlier may be
suggested 62.
Accordingly, the shoreline along the Sea of Marmara must
have possessed permanent harbour facilities as early as the
Roman and possibly the Hellenistic period as shown by the
inscribed spolia from the quay line. Harbour activities prior
to the Byzantine era are indeed attested by a large number
of trading goods and other archaeological objects found
throughout the harbour basin 63. If that should indeed be the
case, contrary to the general perception, the shoreline must
have been quite different as late as the fourth century. This
goes along with Mango’s suggestion of a much deeper bay,
which had only gradually been filled in due to the siltation by
the Lycus river, as well as the continuous land reclamation for
the shaping of the new capital under the reign of emperor
Constantine I and his successors 64.
The jetty and its associated pier construction as well as
the gradual transformation of the coastline provide a rough
terminus post quem for the construction of the quayside,
but the question of its exact date remains. In this regard, the
above-mentioned wooden pier »Marmaray İskele 1« may
shed further light on its historical placement. Running from
the quay in a bow towards the southeast, the 43.5 m long
wooden pier is immediately associated with the coastal fa-
cility. With its four construction phases ranging roughly be-
tween the years shortly after AD 527 and 610 65, it provides
a terminus ante quem of the first half of the sixth century for
the construction of the quay.
As a matter of fact, despite the potentially large time
frame between the Hellenistic, Roman and Early Byzantine
periods, a sixth century date appears to be likely. Considering
the extensive building activities under the reign of Emperor
Justinian I, which included the construction of harbour sites
according to Procopius 66, it is conceivable that the Theodo-
sian harbour underwent some repair or perhaps extension as
well. Although no literary sources refer to any public work of
62 The authors hope that more detailed information on the jetty and its pier pro-
jection will be disclosed and published in future by the Istanbul Archaeological
Museum.
63 Asal, Yenikapı excavations 7; Ercan, Yenikapı 58; Külzer, Harbour of Theodosius
37, in this volume; Öncü, Greek-Roman period.
64 Mango, Shoreline 20-21 fig. 1.
65 Kuniholm et al., Of Harbors and Trees 67-68 fig. 6; it has again to be made
aware that the dating of the wooden remains always refer to the time of their
cutting and not necessarily to their immediate use.
66 Prokopios, De Aedificiis I 8. 1-9 and I 11. 16-20 (IV 33-34. 43-44 Haury / Wirth);
Ercan, Yenikapı 48. 50. 125; Hohlfelder, Building Harbours 369.
67 Prokopios, De Aedificiis V 1. 7-16 (IV 150-152 Haury / Wirth); Koder, Aigaion
Pelagos 287-291; Külzer, Harbour of Theodosius 39, in this volume; Müller,
Getreide 5-11.
68 Kuniholm et al., Of Harbors and Trees 67-68 fig. 6.
69 Pulak / Ingram / Jones, Byzantine Shipwrecks 47-50. Due to its small dimension
(with a documented length of 9 m and a width of 3 m), the likewise 7th-cent.
38 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
such scale and significance, the construction of the granaries
on the island of Tenedos under the reign of Emperor Justin-
ian I must have had quite an impact on the harbour activities
and hence the required infrastructures and harbour installa-
tions 67. Thus, although a late fifth century date or the reigns
of Justinian’s immediate predecessors, Emperor Anastasius I
(491-518) and Justin I (518-527) are equally conceivable for
its erection, the wooden remains of the pier show a perfect
match with that of Justinianic sites such as Capidava 68. As
such, the construction of the quayside has most likely been
implemented as part of the extensive building programme
during the reign of Justinian himself.
The last phase of the wooden pier »Marmaray İskele 1«
shows a continuous use of the western harbour basin at
least up to the mid-seventh century. This is supported by the
shipwreck YK 11, which was unearthed in close vicinity to the
eastern end of the wooden pier (fig. 19) 69. However, the con-
dition of the ship’s hull points to its abandonment in shallow
water. As such, at the time of its dereliction at some point
during the seventh century, the western harbour basin must
have already suffered from heavy siltation by the Lycus River.
Additionally, due to different environmental effects such
as earthquakes during the sixth century, a sudden altera-
tion of the coastline, which, along with the siltation by the
river Lycus, led the harbour basin to become shallower, may
have required the shift of harbour infrastructures or even
the construction of new facilities 70. This correlates well with
the building activities in the eastern harbour basin discussed
above, as well as further historically documented harbour
works such as the construction of the eastern jetty or the
repair and extension of the sea walls around the harbour at
the turn of the seventh to the eighth centuries.
To its west, the quayside is confined by a breakwater of
20 m length, aligning on a northwest-southeast direction
(fig. 30) 71. Similar to the quay line, unfortunately it could
not be uncovered in its entirety. Thus, its structural and func-
tional characteristics can no longer be reconstructed with
certainty. However, some technical and architectural conclu-
sions can still be drawn. Reaching up to the surface of the
quay, the breakwater must have protruded from the surface
of the sea (fig. 31) 72. Consequently, it can be identified as of
YK 11 wreck has been identified as a local cargo vessel for coastal shipping.
Hence, it might form the link between the quayside and the harbour activities
at the western harbour basin of the Theodosian Harbour and the granaries on
the island of Tenedos: Külzer, Harbour of Theodosius 39 n. 48, in this volume.
70 Ercan, Yenikapı 106; Guidoboni, Earthquakes 292-295; Külzer, Harbour of The-
odosius 39, in this volume.
71 Ercan, Yenikapı 86. 135; Gökçay, Architectural Finds 170-171. A breakwa-
ter forms an artificially placed construction, which provides protection to un-
sheltered harbour sites against the prevailing strong sea waves, currents and
tides. By breaking the force of the sea, it assured a safe anchorage for ships:
Dear / Kemp, Ships and the Sea 65; Ginalis, Byzantine Ports 26; Feuser, Hafen-
städte 229-230.
72 Whether the height of its projection from the water allowed waves to break
over it in order to prevent siltation by creating controlled currents within the
harbour basin, remains unanswered. The erection of sea walls points to the
absence of an effective de-silting measure, which resulted in an even faster
siltation of the harbour basin (see below).
type »Mound Breakwater« 73. Typically for a mound-formed
type, the breakwater consists of two different construction
parts: one internal and one external. Built in cross-sections,
it started from the core to the outer protective covering. The
core comprises a mixture of debris or soil with small stones in
order to gain maximum compactness, whereas the external
part consists mainly of larger quarry stones.
The purpose of the external part was mainly to prevent the
movement and washing out of the internal rubble material.
According to the stone size used for the core part as well as
the thickness of the outer covering, sometimes a second layer
of stones was required to cover the whole mound 74. This,
however, cannot be verified here. Its efficiency and stability
depended, not only on the size of the feature, the thickness
of the stones and the weight of the composition, but also on
the gradient of the slope. The slope provided stability for the
construction material by preventing possible undermining by
the sea. The gradient of the slope differs between the inner
and the outer part of the breakwater. While the inner part
(the side towards the harbour basin) drops quite abruptly
with a steep vertical angle, similar to the breakwater at the
harbour of Chrysopolis at Üsküdar (see below), the outer
part (the side towards the open sea) probably possessed a
gentle inclination, which must have started nearly from the
middle of the structure. This provided the construction with
the necessary stability against the strong winds and absorbed
the force of the waves from the open sea. Beyond its struc-
tural composition, the breakwater shows residues of mortar
(most probably again a certain type of hydraulic concrete)
encrusted with the rubble filling of the breakwater’s external
section. This forms a compact mass, which probably acted as
reinforcing binding material for the weight of any superstruc-
ture 75. Indeed, a wide flat surface follows the steep-angled
inner edge, which supports a wall 2.3 m high and 1.35 m
wide (figs 26. 30. 32) 76.
Concerning the construction date of the breakwater, it has
to be noted that the chronological determination of break-
waters turns out to be difficult, since they pertain to a type of
construction that has remained architecturally unchanged for
millennia. As such, a relatively accurate dating often relies on
constructional details, as well as on associated buildings and
archaeological finds, respectively. While its structural char-
acteristics and the use of mortar (hydraulic concrete) allow
a time frame between the Roman Imperial and the Middle
Byzantine periods, the fact that the breakwater is stratigraph-
ically overlapping the quay is certainly decisive for its dating
73 For the construction and typology of breakwaters, see Cornick, Engineering
116. 118 ff; Ginalis, Byzantine Ports 26-31.
74 Ginalis, Byzantine Ports 28, Ill I.7a, vol. II.
75 Gökçay, Architectural Finds 170.
76 Gökçay, Architectural Finds 172.
77 Heher, Harbour of Julian 52, in this volume. It should be mentioned that Heher
wrongly agrees with van Millingen, Walls 291. 294 in the interpretation of the
construction works as a mole. The term προβόλους should rather be inter-
preted as breakwater (see also προβόλιον and προβάλλω in LSJ 1470. 1472).
This is verified by Cristoforo Buondelmonti’s depiction of the harbour and
Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
(fig. 31). Consequently, contrary to the interpretation of the
excavators, the quay provides a terminus post quem of the
late fifth to early sixth century for the construction of the
breakwater and thus also for its wall superstructure. A sixth
century date is also supported by its neighbouring Harbour
of Julian / Sophia (later the Kontoskalion Harbour), for which
the construction of a breakwater under the reign of Emperor
Anastasius I (491-518) is attested 77 (fig. 33).
As for the wall superstructure, according to the excavators
it is supposed to belong to the Theodosian sea wall, forming
its extension along the breakwater 78. However, taking into ac-
count the proposed dating limit by its breakwater foundation,
together with further construction measures to the north and
west of the harbour facilities 79, a date after the mid-sixth
century should be considered. On closer examination, this
wall seems indeed to be architecturally slightly different to the
Theodosian walls surrounding the city. The Theodosian walls
are constructed with a core of mortar faced with carefully
cut limestone blocks and regular bands of brick 80. Although
the wall on the breakwater consists of successive courses of
ashlar blocks with traces of brick bands that recall the Theo-
dosian walls, it shows a much simpler and irregular construc-
tion with building material comprising small to medium-sized
stone blocks and spolia (fig. 30).
Whether the erection of the breakwater and that of its
wall superstructure are to be dated to the same period re-
mains speculative. This question is closely related to the yet
unexplained function of a series of holes drilled below the
wall in an east-west direction. This continuous row of holes
would appear to indicate that wooden beams connected
the breakwater conglomerate to the superstructure (figs 30.
32) 81. Traces of mortar coating suggest that at least the lower
part of the wall and the holes were plastered. This would
have protected the wooden features, which easily deform,
swell or decompose when in contact with water. Unfortu-
nately, it is still unknown whether the drilled wooden beams
were only intended to provide greater stability for the wall
or whether they functioned as connecting elements for the
mortar binding. From an engineering perspective, however,
this building measure is probably best explained as a binding
element for a later, additional construction.
Consequently, it can be suggested that the two features
most likely belong to different construction phases. This argu-
ment is supported by a short stretch of further wall just west
of the sea wall (figs 25-26). Despite a slightly different orien-
tation, these wall remnants may represent a potential sea wall
later by the harbour reconstruction of Müller-Wiener: Effenberger, Pictorial
Sources figs 2. 4, in this volume; Heher, Harbour of Julian fig. 7; Müller-Wie-
ner, Häfen 37.
78 Gökçay, Architectural Finds 170; Kocabaş, Theodosian Harbour 25.
79 Some wall remains of the building complex associated with the harbour instal-
lation revealed stamped bricks dated to the 6th c. in situ: Ercan, Yenikapı 114.
80 For the Theodosian walls of Constantinople, see Asutay-Effenberger, Land-
mauer 13-71; Mango / Kiefer / Loerke, Monuments 519-520; Turnbell, Walls;
van Millingen, Walls.
81 Gökçay, Architectural Finds 171.
39
as well, belonging to the initial phase of the breakwater and
thus predating the sea wall discussed above. These eventually
served as a supporting structure for the later sea wall with its
buttresses that is visible today (fig. 26). Consequently, it can
be proposed that the breakwater initially carried a different
sea wall, which was erected, together with its substructure,
probably around the sixth century (either together with the
quay or slightly later). Subsequently, it must have been re-
placed by the above-discussed sea wall at some later point,
postdating the entire building complex. According to written
accounts, the sea walls supposedly underwent repairs either
during the reign of Tiberius II (698-705) or Anastasius II (713-
715) 82.
In fact, an identical building activity with identical con-
struction phases can be observed at the Byzantine harbour of
Thessaloniki, where rescue excavations unearthed a section of
the sea wall with successive courses of ashlar blocks having
traces of brick bands and buttresses 83. Similar to its equivalent
at the Theodosian harbour, based on the archaeological anal-
ysis and the written account of Caminiates, at least two major
construction phases have been determined for the Byzantine
era 84. While it is suggested that the earlier one is dated to
the Early Byzantine period, the following construction phase
has to be placed at some point between the mid-seventh and
the second half of the ninth centuries. Hence, similar to the
inner circuit walls enclosing the harbour basin of the Theo-
dosian harbour (fig. 34), a construction date at the turn of
the seventh to the eighth centuries, may also be considered
for the later sea wall 85. However, only a closer investigation
of the building material at both harbour sites, such as the
mortar filling or the incorporated bricks, will provide accurate
dating information.
In conclusion, the archaeological excavations at Yenikapı
revealed a nearly complete historical sequence of human
activities in the Theodosian harbour, ranging from its foun-
dation in the late fourth or early fifth centuries (if not al-
ready from the pre-Byzantine era) up to its final rededication
in the fifteenth century. This provides not only information
on traded goods and artefacts in daily life, but also much
sought-after information on shipbuilding traditions, as well
as on harbour installations and their architecture from Late
Antiquity to the Late Middle Ages. Interestingly, as far as the
physical remains of coastal facilities are concerned, they nicely
demonstrate the evolution of harbour activities. The earliest
infrastructures, which belong to the first construction phase
of the harbour around the end of the fourth or the beginning
82 Külzer, Harbour of Theodosius 40; Müller-Wiener, Bildlexikon 313.
83 Leivadioti, Thessaloniki 87, Eικ. 49α-β.
84 Kameniates, De expugnatione VIII 3 (9 Böhlig); ibid. 22-25. It should further
be noted that pre-Byzantine building remains have been discovered as well,
indicating an earlier construction phase dating to the Roman period: Leivadioti,
Thessaloniki 20-21.
85 Dark, Post Office Site 318; Külzer, Harbour of Theodosius 40; Mango, Shoreline
24-25; Müller-Wiener, Häfen 9.
86 Ercan, Yenikapı 115-116; Külzer, Harbour of Theodosius 45-46.
40 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
of the fifth centuries, were erected along the northern coast
of the harbour basin just west of the outlet of the Lycus River
(fig. 3). These include a jetty (the western jetty) and wooden
piers. The wreck finds of YK 22, YK 26, YK 34, YK 35 and
YK 37 (figs 2. 19-20), together with stamped bricks dumped
next to the piers (fig. 35), attest to the loading and unloading
of traded goods as early as the fifth century 86.
Based on the distribution of commercial installations listed
in the fifth-century Notitia urbis Constantinopolitanae 87, it
can be assumed that some harbour infrastructure may also
have existed east of the river estuary (fig. 36). Accordingly,
granaries (so-called Horrea) and marketplaces (so-called Fora)
in its northern and eastern periphery, such as the Horrea Al-
exandrina, the Horrea Theodosiana, the Forum of Theodosius
or the Forum of Amastrianus, show a close connection to the
Harbour of Theodosius 88. Although the grooved ashlar blocks
on the eastern jetty may have belonged to a potential quay
structure along the eastern end of the harbour, further phys-
ical remains of harbour facilities have yet to be discovered.
Concerning the western harbour basin, the current state of
archaeological knowledge suggests that no harbour installa-
tion existed at that time.
However, the pre-Byzantine building material of the west-
ern quayside allows the assumption that an earlier harbour
site, possibly from the Hellenistic or Roman periods, may
have existed. If so, this must have been situated slightly far-
ther north or north-west and was later dismantled for the
construction of the new quayside in the Early Byzantine pe-
riod. Whether its facilities had been in use until the reign of
Emperor Constantine I (306-337) remains unknown for the
time being.
A wall, approximately 54 m long and 4.40 m wide, was un-
earthed just north-west of the quay and breakwater (figs 26.
37). Running beneath the sea wall discussed above, the exca-
vators speculated that this 1.9 m high wall fragment formed
the southern limit of the so-called »Constantinian« wall 89.
The latter reached the shoreline of the Sea of Marmara further
west, implying that a certain section of the wall ran along
the coast. According to Mango, however, the sea walls could
hardly have existed under the reign of Constantine I due to
the steady change of the coastline, as well as the continuous
land reclamation of the deep bay (which later became the
Harbour of Theodosius) up until the very end of the fourth
century 90. Only by AD 439, at the earliest, the construction
of maritime circuit walls was finally ordered. At any rate, it
seems that new harbour construction works on this shore
87 Notitia Urbis Constantinopolitanae 237. 239 (Seeck); Mundell Mango, Com-
mercial Map.
88 Ercan, Yenikapı 21. 59. 65. 78; Heher, Harbour of Julian 52, in this volume;
Külzer, Harbour of Theodosius 39, in this volume; Mundell Mango, Commercial
Map 192-193 fig. 4.
89 Ercan, Yenikapı 110-111; Gökçay, Architectural Finds 172.
90 Mango, Shoreline 18-24; Many scholars accept the fact that the majority of
the buildings attributed to Constantine I could not have been completed under
his reign, but during the reign of his son Constantius II. The wall discussion
basically relates to this debate: Magdalino, Maritime Neighbourhoods.
were not undertaken at least before the end of the fifth
century.
Consequently, the Harbour of Theodosius was probably
extended towards the west at the time of the »reconquest«
and annexation of North Africa under the reign of Emperor
Justinian  I, which opened new markets and trading con-
nections for Constantinople. It must have been that time
when the previous harbour installation was abandoned and
eventually partly removed to be used for the construction of
the new quayside 91. Therefore, it is in the sixth century that
the Harbour of Theodosius, not only experienced its most
prosperous time, but also seems to have reached its largest
extent and final face.
The use of the entire harbour area did not last for long.
The last phase of the wooden pier »Marmaray Iskele 1«,
together with the shipwreck YK 11, demonstrate that the
western harbour basin remained in use only until the end of
the seventh century or beginning of the eighth century. This
was caused by the Lycus River and a series of other environ-
mental effects, which led to a slow but constant siltation of
the harbour basin. The archaeological data obtained from the
harbour installations, as well as the distribution and dating
of the wreck finds, show that the siltation process took place
from west to east 92.
Due to the loss of Egypt, Palestine and Syria to the Arabs
in the seventh century, it was obviously no longer necessary
to use the harbour at full capacity 93. Therefore, in contrast to
the neighbouring Harbour of Julian / Sophia 94, costly dredging
works were not undertaken. Despite the reduction in the size
of the harbour area and thus also the restriction of harbour
activities, the Harbour of Theodosius continued to be a major
hub for maritime trade throughout the Middle Byzantine pe-
riod. The consequences of the Arab conquest of Egypt, which
was the breadbasket of Constantinople, eventually entailed
the reconfirmation of Byzantine authority over the Greek
peninsula in the second half of the seventh century 95. Thus,
while the facilities at the western end of the harbour were
abandoned after all, new infrastructure was constructed in
the eastern harbour basin. Accordingly, in order to meet the
new requirements, a massive eastern jetty was constructed in
close vicinity to the newly renamed granary of Lamia at the
turn of the seventh to the eighth centuries.
At some point after the ninth century, the siltation process
reached the eastern harbour basin with the water depth
constantly dropping. The retreat of the navigable sea level
91 The re-use of construction material from preceeding harbour installations is also
suggested for the harbour of Thessaloniki: Leivadioti, Thessaloniki 21.
92 Ercan, Yenikapı 135; Külzer, Harbour of Theodosius 41, in this volume.
93 Of course, one should not ignore the impact of the decline in population due
to famine and pestilence: Stathakopoulos, Famine and Pestilence.
94 Heher, Harbour of Julian 52-53, in this volume.
95 Ginalis, Byzantine Ports 238-239; Trombley, Boeotia 991-992. Contrary:
Koder / Hild, Hellas und Thessalia; Lilie, »Thrakien« und »Thrakesion« 35-41;
Haldon, Palgrave Atlas. – For the Arab conquests see: Kaegi, Early Islamic Con-
quests.
96 Effenberger, Pictorial Sources 20 fig. 1, in this volume; Ercan, Yenikapı 62. 92.
96. 118; Kocabaş, Theodosian Harbour 32; Külzer, Theodosius-Hafen 41-42;
Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
towards the south by the end of the twelfth century, at the
latest, eventually required further building measures at the
eastern end of the eastern harbour basin in order to reach the
necessary draught for docking vessels. However, the exten-
sion of the eastern jetty towards the south shows that these
may no longer have comprised major and elaborate harbour
constructions. While the wreck sites attest to the use of the
Harbour of Theodosius only up until the end of the Middle
Byzantine period, some Late Byzantine kilns, together with
written sources and depictions, point to minor harbour activ-
ities up to the fifteenth century 96.
Finally, based on the archaeological study of the Harbour
of Theodosius, a very last observation is worth mentioning.
The above-discussed constant siltation process of the harbour
basin is not just associated with the Lycus River and a series of
other environmental effects, as well as human impact (e. g.,
by dumping waste material into the harbour). Additionally,
sea currents passing through the harbour mouth had a con-
siderable effect. Consequently, the unequivocal west-east
shift of the siltation process is closely related to the angle of
the confluence between the incoming currents and the river
outflow (fig. 38). Thus, the position and direction of the
harbour mouth must have played a decisive role.
Accordingly, this may provide an indication for a potential
reconstruction of the orientation of the breakwaters and its
sea wall superstructures. Most recent harbour reconstructions
suggest two equal breakwaters with a centrally located har-
bour entrance in an eastern orientated direction (fig. 39) 97. In
order to perform the west-east effect on the siltation process,
the harbour entrance must have been located on the eastern
side (fig. 40). This is also indicated by the different courses
of the two breakwaters. Based on the aforementioned re-
construction models, the western breakwater ran in a more
or less straight east-west direction, whereas the eastern one
showed first a clearly north-south orientation before turning
west where it continued to the centre of the harbour basin.
In line with Mango’s and Janin’s suggestions 98, it should
rather be assumed that only one long western breakwater
existed, which formed a large eastern harbour entrance. The
use of a single breakwater is supported by historical depic-
tions, such as Buondelmonti’s drawing of Constantinople in
his Liber insularum archipelagi, and also by the Byzantine
harbour of Thessaloniki 99. In fact, the same building tech-
nique can also be seen in photographs of the other harbour
sites along the coast of the Sea of Marmara, the Harbour
Magdalino, Maritime Neighbourhoods 215. Although it is beyond the scope of
this article, it is important to note that at an elevation nearly equal to the later
kilns, a small church was constructed to the southeast of the jetty possibly after
the 10th or 11th c. For the archaeological analysis of the church and the theory
about its abandonment in the 13th c., see Gökçay, Architectural Finds 166-180;
Ercan, Yenikapı 80-82; Marinis, Architecture 208.
97 Berger, Langa Bostanı figs 1-4; http://www.byzantium1200.com/port_t.html (5
February 2020).
98 Janin, Constantinople Map 1; Mango, Shoreline fig. 1.
99 Effenberger, Pictorial Sources figs 1-2, in this volume; Leivadioti, Thessaloniki
Σχεδ. 2.
41
of Julian / Sophia, the Boukoleon harbour and the mooring
areas at Hebdomon (modern Bakırköy) and Brachialion (see
below) 100.
Interestingly, all the harbours along the Sea of Marmara
coast obviously possessed one single breakwater coming
from the west. Exactly the same orientation of entrances is
also shown by the siting of the modern harbours. All the har-
bour sites facing the Sea of Marmara possess just one single
breakwater coming from the west, thus forming an eastern
harbour entrance. With a southeast to east direction, they
enclose and protect the harbour basins against the prevailing
south-western and southern winds 101. At the same time, the
south-eastern to eastern currents must have made entrance
into the harbour basins easier and also acted as a natural
measure against their siltation during the Byzantine era.
Further Remarks on the Physical Remains
of the Harbours of Constantinople
and its Hinterland
Regarding the physical remains of harbour installations at
other coastal sites in Constantinople, archaeological inves-
tigations were recently carried out at the harbours of Chal-
cedon at Kadıköy and that of Neorion / Prosphorion 102 at
Sirkeci 103. While the salvage excavation at Kadıköy supposedly
revealed the remains of a jetty using a hydraulic concrete
base similar to that found at the Harbour of Theodosius and
Chrysopolis, a number of architectural elements and plenty of
pottery finds belonging to the Early to Late Byzantine periods
were brought to light at Sirkeci. These architectural elements
comprise wooden structures, possibly connected with the
harbour’s surrounding warehouse facilities 104.
Within the archaeological context of the wider harbour
bay, an eyewitness report by Charles Marling from 1906
provides vague but important information on further har-
bour features within the Neorion / Prosphorion Harbour 105.
According to his letter to Arthur B. Skinner, he observed a
row of stone blocks of around 1 m in dimension during rescue
excavations at the new post office south-east of the Ottoman
Spice Bazaar, which he interpreted as a quay structure 106. The
method of construction, the use of building material and
geological and ceramic evidence led Dark to support an Early
Byzantine (fourth to seventh centuries) date 107. In fact, con-
sidering the characteristics of quay structures at other harbour
100 Heher, Harbour of Julian 54. 63-64, in this volume; Heher, Bukoleon 67 fig. 5,
in this volume; Simeonov, Hebdomon 127, in this volume; Simeonov, Brachi-
alion 139, in this volume.
101 Heher, Boukoleonhafen 133.
102 The division of the wider bay along the northern coast of the peninsula and
hence the exact location and separation of the two harbours is still uncertain:
Kislinger, Neorion, in this volume; Dark, Harbours 153-154; Dark, Post Office
Site 317.
103 Kızıltan, Yenikapı, Sirkeci and Üsküdar 15-16. The archaeological works at
Kadıköy have not yet been published.
104 See Mundell Mango, Commercial Map 200-201 fig. 4; Kislinger, Neorion 94
n. 42, in this volume.
42 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
sites, an earlier date of the Roman, if not even Hellenistic,
period may equally be suggested.
Similar data has been revealed by the Sirkeci Station Res-
cue Excavation between 2004 and 2012. Just as Dark sug-
gested for the site at the new post office, a large building
complex of the fifth to seventh centuries was documented
at the eastern shaft of the Sirkeci Metro Station (north of
the train station) 108. After removing the Early Byzantine
strata, however, wooden structures and so-called »water-
front stones« were revealed 109. Together with wooden ships
remains, these seem to belong to the harbour facility of the
Prosphorion harbour itself. Unfortunately, no further data
has yet been published, which could provide more detailed
information. Only a single photograph offers a first glimpse
of the so-called »waterfront stones« after all.
It shows a semi-circular three-levelled row of stones, which
can be identified as a quayside (fig. 41). Each row is offset by
around 20 cm and consist of roughly 1.20 m × 0.90 m large
ashlar blocks. Among the almost uniform rectangular ashlar,
at least two stone blocks possess lifting bosses for their place-
ment on site. Such construction measures are mostly known
from harbour sites of the Hellenistic period like the harbour
of Amathus on Cyprus 110. Archaeological finds dating as early
as the seventh century BC have indeed been documented
during the rescue excavation. However, since the quay line
only includes a very small number of stone blocks with lifting
bosses, which also seem to have been placed randomly, it
can be assumed that these form reused material from a pos-
sible earlier harbour installation. The fact that metal clamps
characteristic for classical antiquity are missing as well further
supports a post-Hellenistic date. On the other hand, no mor-
tar binding material seems to have been used either, which
gives the quay construction an isodomic character. As such,
given the use of hydraulic concrete for the Early Byzantine
harbour facilities at Yenikapı, a date to the Roman period may
rather be suggested. This is also indicated by its architectural
characteristics, with the three-stepped construction method
finding parallels both in Hellenistic and Roman harbours such
as Mytilene, Leptis Magna or the river quay of the Tiber in
Rome 111. Finally, remains of two marble columns can be ob-
served just next to the quayside. Although they seem to be
of later, possibly even Early Byzantine date, they appear to be
aligned with the quay. As such, the columns either belonged
to an associated building, or formed mooring facilities for
berthing ships.
105 Kislinger, Neorion 93. 95 fig. 2; Dark, Post Office Site 315.
106 Unfortunately, the archaeological structure has never been published.
107 Dark, Post Office Site 317-318.
108 Gür, Rescue Excavations 17; Gür / Emre, Sirkeci 32-33.
109 The architectural interpretation of the remains will be subject of examination
by K. Gür within the scope of an ongoing doctoral dissertation at Istanbul
Technical University. For preliminary results, see Gür, Rescue Excavations 16-
17; Gür / Emre, Sirkeci 32-33; Kızıltan, İstanbul Kazıları 364.
110 Empereur et al., Amathus 62-65.
111 Blackman, Ancient Harbours II 203 fig. 11; Ginalis, Byzantine Ports 34. 40;
Theodoulou / Kourtzellis, Lesbos Underwater 97. 99.
 Whether the harbour structures unearthed at the Sirkeci
Station and those observed by Charles Marling during the
rescue excavations at the new post office belong to the same
harbour installation remains unclear. Considering the almost
identical dimensions given for the stone blocks, it is quite
feasible to assume it though. Far more convincing that the
two sites may actually reflect parts of the eastern and western
areas of one and the same infrastructure is their distance of
only 300 m as well as their identical height in relation to the
shoreline 112.
However, it is entirely possible that one belongs to the
Neorion and the other to the Prosphorion harbour as well.
Whatsoever, permanent harbour infrastructures at the Ne-
orion / Prosphorion Harbour can be traced back at least to
Hellenistic times with the current facility most likely repre-
senting the Roman phase by reusing material from the earlier
harbour installation. A continuous use of both quay lines up
to the Byzantine era has yet to be ascertained but seems likely.
In any event, as one of the most important and most fre-
quented harbours of Constantinople, the detailed analysis of
the harbour remains of Neorion / Prosphorion as well as that
of Chalcedon will provide new ground-breaking information
for harbour studies of the pre-Byzantine, Early Byzantine and
Late Byzantine periods.
For the rest of the Constantinopolitan harbours, infor-
mation is even more limited. In the case of the Harbour of
Julian / Sophia, the sources of information comprise only spo-
radic and rough drawings 113. For the Harbour of Julian / So-
phia (fig. 33), as well as for the mooring areas at Hebdomon
at Bakırköy (fig. 42) and Brachialion at Mermer Kule (fig. 43),
one can rely at least on a few photographs. Nevertheless,
the drawings and photographs generally provide only rough
impressions of the various harbour infrastructures. The only
exceptions are the harbours of the Boukoleon Palace and
Chrysopolis at Üsküdar.
The Boukoleon Harbour
The intensive studies of the Boukoleon Palace and its sur-
rounding sea walls entailed a more detailed photographic
documentation of at least part of its harbour facilities (fig. 44).
As for the latter, Heher rightly assumes that, with the exten-
sion of the Great Palace towards the south, the first mooring
facilities must have existed as early as the time of Emperor
Justinian I 114. This is supported by Procopius’ reference to the
anchoring of Belisarius in front of the palace, as well as by
112 Kislinger, Neorion 93 fig. 2, in this volume.
113 See Heher, Harbour of Julian figs 2. 7. 9, in this volume.
114 See Heher, Harbour of the Bukoleon 70-71; Heher, Boukoleonhafen 123. 125.
115 Prokopios, Bella III 12. 2 (I 365 Haury / Wirth); Heher, Harbour of the Bou-
koleon 71, in this volume; Özgümüş, Bukoleon 66.
116 Mamboury / Wiegand, Kaiserpaläste 13.
117 Ginalis, Anthedon; Schäfer, Larymna 533-537 fig. 14; Schläger / Blackman / 
Schäfer, Anthedon 36 figs 9. 14.
Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
ceramic finds and the sea walls in this section, which show a
construction phase of the sixth century 115. The photographic
documentation of the harbour mainly includes a quayside
at the eastern end of the harbour (fig. 45) that runs in front
of Justinian’s house (leading from the grand staircase to the
lighthouse tower) (fig. 46). Fortunately, the high quality of
Mamboury and Wiegand’s photographic record still allows
a clear recognition of the quay structure. Accordingly, it is
clearly visible that it consisted of massive limestone ashlar and
large marble blocks (0.6 m × 0.7 m), according to Mamboury
and Wiegand having a total width of at least 6 m and at the
grand staircase even up to 12 m 116.
On closer examination, it can be observed that not all of
the blocks show an identical orientation. While the quay is
seemingly constructed of rows of east-west-running stone
blocks, north-south-running rows of ashlar were inserted
at regular intervals (figs 47-48). This gives the impression
of a chamber system, among others strongly resembling
the quaysides of the harbours of Anthedon and Larymna
(see below) 117. The chambers must have been filled with a
type of hydraulic concrete, consisting of a conglomerate of
rubble stones and mortar with inclusions of coarse ceramic.
These were subsequently covered with the limestone ashlar
blocks and with marble blocks around the grand staircase,
as this has been nicely reconstructed by Helbert (fig. 49) 118.
Anna Komnene claimed that the harbour had been built
using mortared fieldstones and marble blocks 119. This is also
indicated by Mamboury and Wiegand’s report of large lime-
stone blocks, quarry stones and brick mortar 120. Recent core
drillings in front of Justinian’s house and the southeast cor-
ner of the grand staircase further attested this construction
method 121. At a depth of approximately 3.75-4.8 m, the
latter revealed an artificial conglomerate of clayey sand with
brick and stone, as well as marble fragments. This is fol-
lowed by a stratum of mudstone, made of dark grey stones
between 4.8 m and 6.9 m, and finally a layer of brown-grey
gravel / rubble stones and clayey sand with brick inclusions
down to a depth of 10 m. While the top layer (approximately
1 m thick) obviously represents the cover plates of ashlar and
marble blocks with their mortar binding, the following lay-
ers most likely form the compact mortar filling of the quay
chambers.
Similar structural remains have also been documented
around 50 m east of the so-called Tower of Belisarius and
40 m south of the façade of the western palace section
(fig. 46). Mamboury and Wiegand referred to an »isolated
foundation of quarry stones with brick mortar, followed by
118 The marble blocks have most likely been re-used, as this is the case also for
the use of marble spolia for the construction of the sea wall’s lower section:
Mango, Boukoleon 47.
119 Heher, Boukoleonhafen 133; Heher, Harbour of the Bukoleon 80.
120 Mamboury / Wiegand, Kaiserpaläste 6. 13.
121 Bolognesi Recchi Franceschini, Monumental Itinerary 55-56.
43
four layers of bricks«, which they interpreted as the founda-
tion of the eponymous animal statue »Boukoleon« 122. Heher
rightly doubts this interpretation and in turn proposes an
identification as a quay line belonging to a large ceremonial
square. Alternatively, he suggests that it may have been part
of a breakwater or mole construction 123.
Given the description of the structural remains, the feature
obviously constituted a hydraulic concrete foundation with
a brick superstructure. As such, it can indeed be identified
as the architectural element of some sort of harbour infra-
structure. The proposed existence of a quay-like facility that
featured a ceremonial area is therefore conceivable. However,
given the location of the remains, it seems more likely to
assume a mole construction on the inner side of a break-
water, which enclosed the harbour basin coming from the
western sea wall or the Tower of Belisar 124. This would not
only explain Buondelmonti’s drawing of an enclosed harbour
basin 125, but also agree with Nicetas Choniates’s statement
»…περὶ τὰς ἀκτὰς σαλεῦον καὶ τοὺς προβλῆτας, οἳ τὸ πάραλον
τεῖχος τῆς πόλεως διειλήφασι, τὰς τῶν κυμάτων ἀποθραύοντες
ἐμβολάς« 126.
Concerning its superstructure, it seems that the brickwork
rests directly on the concrete foundation. As such and in con-
trast to the eastern jetty at the Harbour of Theodosius, we
might find here a continuation of the Roman tradition com-
bining concrete with brick for the construction of an arched
mole structure. Given the arches or blind arcades at the
western sea wall, the existence of an arched mole seems not
too far-fetched. Whether the latter formed just blind arcades
as decorative elements or proper arches remains unknown.
In fact, in order to find a way to act against the problem of
siltation in a small harbour like the harbour of the Boukoleon
Palace, such a building measure would have certainly made
sense. While the breakwater substructure reduced the force
of the waves and hence broke the strength of the sea, it al-
lowed the waves to break over it. Passing through the arches
of the mole, these subsequently created currents within the
harbour basin 127.
As for the dating of the quay construction at the eastern
end of the harbour, the chamber system of intersecting lateral
and longitudinal walls finds comparison in the eastern jetty at
the Harbour of Theodosius, as well as in quaysides and jetty or
mole constructions of the seventh to eighth centuries at a se-
ries of harbour sites, such as Anthedon, Larymna, Theologos,
Aegina, Thessalian Thebes and Lechaion 128. Consequently,
122 Heher, Boukoleonhafen 134; Mamboury / Wiegand, Kaiserpaläste 5, tab. VII,
XXXV.
123 See Heher, Harbour of the Bukoleon 82, in this volume; Heher, Boukoleon-
hafen 135.
124 Even though the structural remains could have easily belonged to a jetty as
well, such identification has to be ignored. Even with a calculated quayside
of around 9 m along the eastern harbour side, the distance of at least 40 m to
the Tower of Belisar is far too great for a jetty in this harbour.
125 Effenberger, Pictorial Sources fig. 1, in this volume.
126 Niketas Choniates, Historia 129 (van Dieten); Heher, Harbour of the Bukoleon
80.
44 Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
the quayside at the Boukoleon harbour seems to again rep-
resent Middle Byzantine harbour architecture.
A Middle Byzantine date for the harbour has also been
suggested by Mango and Heher 129. As opposed to Heher’s
assumption that the quayside could only have been con-
structed after the third and last construction phase of the
sea wall during the ninth to tenth centuries, a date around
the turn of the seventh to the eighth century should rather
be accepted. This matches perfectly with the erection of the
second construction phase of the sea wall and the Tower of
Belisar, as well as that of the grand staircase as a monumental
access to the palace during the reign of Justinian  II (685-
695 and 705-711) or Tiberius III (698-705) 130. An additional
3.2 m was added to the 6 m wide sea wall during the third
construction phase, meaning that it can be assumed that the
quayside originally had a total width of 9.2 m. Unsurprisingly,
this coincides exactly with the width of the seventh-to-eighth-
century-dated southern quay at the harbour of Anthedon and
the eastern quayside at Larymna (for which a width of 4.6 m
is given for a single chamber) 131. Accordingly, the quayside
would originally have had a double-chamber construction.
As a result, Mango is right in assuming that an independ-
ent palace harbour approximately 1.45 ha in size, with a
harbour basin of up to 250 m long and 40 m wide, intended
for private imperial use, was constructed, or simply altered
to its final shape, at some point after the sixth century and
before the ninth century 132. Based on the analysis of the ar-
chitectural remains of the various harbour features, a slightly
different picture than that presented by Helbert (fig. 50) can
be suggested: a significant difference may be proposed for
the western harbour basin. While the existence of a quayside
along the entire façade of the Boukoleon Palace can indeed
be assumed, the harbour cannot have reached as far south
as the Tower of Belisar. The breakwater, with its mole super-
structure, should rather be considered as an extension of the
western sea wall. As such, it can be doubted whether the
palace harbour ever featured a supposed ceremonial square.
Accordingly, it can be suggested that, with the new har-
bour situation in the Middle Byzantine period, the main em-
barkation and disembarkation area even shifted from the
western to the eastern harbour basin. This is supported by
the massive quayside in front of Justinian’s house (figs 47-48)
and by the change of the access point to the Boukoleon Pal-
ace. While the first mooring facilities were accessible through
a 2.7 m wide gate at the western harbour basin during the
127 Ginalis, Byzantine Ports 31.
128 Ginalis, Anthedon; Ginalis, Byzantine Ports 190; Knoblauch, Ägina 73; Paris,
Lechaion 10-11; Rothaus, Lechaion 295-296; Schäfer, Larymna 533-537
fig. 14; Schläger / Blackman / Schäfer, Anthedon 36 figs 9. 14; Triantafillidis / 
Kou­tsoumba, Aegina 169.
129 See Heher, Harbour of the Bukoleon 79, in this volume; Mango, Boukoleon
47.
130 Bolognesi Recchi Franceschini, Seventh Survey 137-138; Heher, Boukoleon-
hafen 126. 129; Heher, Harbour of the Bukoleon 73; Mango, Boukoleon 47.
131 Ginalis, Anthedon; Schäfer, Larymna 533.
132 Heher, Boukoleonhafen 132-135; Mango, Boukoleon 47.
reign of Justinian I, in correspondence with the alteration of
the harbour at the turn of the seventh century to the eighth
century the construction of the second phase of the sea wall
shifted the gate to the eastern harbour basin 133. This seems
also to have been taken into account in the later construc-
tion of the grand staircase, which shows a large eastern gate
(figs 44b-d. 51) 134.
The Harbour of Chrysopolis at Üsküdar
Originally forming a deep sheltered bay, the harbour site of
Chrysopolis at Üsküdar was used as a strategic interstation
for the shipping lane through the Bosporus since classical
antiquity 135. Alongside the coastal sites of Chalcedon at
Kadıköy, Hiereia (Hieron) and Eutropiu Limen at Kalamış bay,
it additionally acted as an important ferry harbour linking
Constantinople with its Asian coast opposite. Furthermore, it
formed a so-called Epineion 136 for Bithynia from the Roman
Imperial period onwards and especially during the Byzantine
era 137. Hence, it is not surprising that again plenty of archae-
ological evidence of harbour activities were brought to light
during the Marmaray-Metro Construction Project between
2004 and 2008 138. The salvage excavations revealed, not
only a large number of ceramic artefacts, marble objects,
stone anchors and a variety of architectural elements rang-
ing from pre-Classical times to the Ottoman period, but also
various building remains belonging to harbour infrastruc-
tures 139. These include a breakwater with a possible mole
construction 140, jetties, wooden pier remains and a potential
quayside 141. The harbour situation is again quite complex
due to the strong alteration of the coastline. The harbour site
of Chrysopolis must have shifted quite frequently through
time as a result of the constant regression of the bay due to
siltation from the estuaries of the rivers Bülbül and Çavuş 142.
Therefore, it must be assumed that the harbour sites of the
Archaic, Classical, Roman and Byzantine periods are situated
in different locations. Similar harbour situations are also doc-
umented at other coastal sites in Asia Minor such as Ephesus
or Clazomenae 143.
As for the unearthed harbour structures, the most striking
features again constitute wooden remains, which belong to
133 Heher, Boukoleonhafen 129; Heher, Harbour of the Bukoleon 75-77 fig. 21.
134 Mamboury / Wiegand, Kaiserpaläste tab. XXIII.
135 Karagöz, Khrysopolis Liman 401. 404. 414; Karagöz, Khrysopolis – Scutari 3
fig. 7; Karagöz, Excavations 86.
136 During classical antiquity, the epineion (ἐπίνειον) constituted a harbour area
outside its associated city, but yet forming a part of it. During the Roman
Imperial period, these so-called out-ports developed into independent coastal
sites, often taking over the role and significance of their preceding ancient
cities. The latest by the Early Byzantine period epineia formed crucial coastal
centres, which acted as vital economic hubs and linking stations for the settle-
ment network within a certain province: Ginalis, Byzantine Ports 15. 250-252.
137 Belke, Bithynien und Hellespont 296-298; Belke, Gates 166, in this volume;
Karagöz, Khrysopolis Liman 406. 412.
138 Belke, Gates 165; Karagöz, Excavations 85; Kızıltan, Yenikapı, Sirkeci and
Üsküdar 15.
139 Karagöz, Excavations 89-101.
Archaeology of the Harbours of Constantinople  |  Alkiviadis Ginalis  ·  Ayşe Ercan Kydonakis
various sections of pier constructions, as well as to jetty form-
works or caissons. Although the wooden pier structures could
not be entirely excavated, remains with a dimension of 8.4 x
4 m and even as large as 8.2 x 5.25 m have been uncovered 144.
Unlike the equivalent remains documented at the Harbour of
Theodosius at Yenikapı, the piers do not consist of vertical
piles only. For the first time, horizontal grid systems with a
floor level have also been preserved, which provide an unique
insight into the engineering details of pier constructions as
shown for example by the depiction of Gregory of Nazianzus’
departure from Constantinople dated to the eleventh century
(fig. 52) 145. The horizontal grid consists of 0.25 m wide and
5.2 m long carved wooden logs set in a grid on top of each
other and pegged on the vertical piles driven into the ground
(fig. 53). A mortise-and-tenon joinery system was applied for
the fastening of the wooden elements 146. One pier section
indicates that the uppermost layer of logs corresponds to the
orientation of the pier. This last layer of logs was subsequently
covered with planks, thus set perpendicular to the orientation
of the pier (fig. 54). As for the construction material itself, it
seems that the timbers were fired and pitched in order to
provide longer resistance to deterioration in the maritime
environment 147.
The pier sections possess a roughly northwest-southeast
orientation. Interestingly, its various parts show a different
river sediment infill. While the north-western section is filled
with pure sand, at the south-eastern end boulders can also
be found. Whether the latter derive from the siltation pro-
cess or whether they were set in order to reinforce the pier
in connection to another harbour structure and support it
against environmental impact, respectively, remains to be
clarified. Corresponding to an alignment perpendicular to the
north-west oriented shoreline, it points to the fact that at the
time of the construction of the piers, the coastline must have
considerably shifted (up to 1 km) towards the north-west
opening of the deep bay 148.
This shift of the coastline also altered its physical condi-
tion 149. Thus, the change from a deep sheltered bay to an ex-
posed open shoreline eventually required building measures
for the protection of the harbour site. Accordingly, a break-
water had been erected. Like the breakwater at the Harbour
of Theodosius (see above), the composition consists of large
140 A mole forms a masonry structure along the inner side of the breakwater. This
increases the mooring space for the loading and unloading of ships within the
harbour basin in order to extend the commercial and traffic-related functions
of the quay: Ginalis, Byzantine Ports 26. 30; Feuser, Hafenstädte 229.
141 Karagöz, Khrysopolis Liman 402. 408-414; Karagöz, Chrysopolis 46-49. 52.
142 Belke, Gates 165-166; Karagöz, Khrysopolis – Scutari fig. 7; Karagöz, Excava-
tions 101.
143 Ersoy, Clazomenae 2-6; Steskal, Ephesos 327.
144 Karagöz, Khrysopolis Liman 408-410; Karagöz, Yapı 422.
145 Cod. Taphou 14, f. 265r; Aidoni et al., Seaports 21 fig. 5.
146 Karagöz, Chrysopolis 49-50; Karagöz, Excavations 101.
147 Karagöz, Yapı 422.
148 Belke, Gates 165; Karagöz, Chrysopolis 46; Karagöz, Khrysopolis – Scutari 5.
149 Physical conditions indicate the consistence and configuration of a specific
coastline, which is affected by the predominating waves, currents, tides and
winds: Ginalis, Byzantine Ports 9; Karmon, Components 1.
45
quarry stones piled up on top of an internal core of rubble
material 150. In order to provide the construction with stability
against strong winds and absorb the force of the waves to
prevent a possible undermining, the structure shows an in-
clination towards the sea (fig. 55). In contrast to the sloping
outer part of the breakwater, the inner part drops abruptly
with a steep vertical angle. The upper part of the breakwater
is flattened, consisting of cut stones.
A row of limestone blocks ranging in size from 0.5-2.9 m
× 1-2 m × 0.25-0.9 m, with inserted spolia of marble column
fragments from the fifth to sixth centuries, may also be allo-
cated to a mole construction at the steep-angled inner edge
of a breakwater (fig. 56). The existence of a mole is further
supported by traces of intensive loading and unloading ac-
tivities of traded goods attested to by numerous amphora
fragments 151. After removing the limestone ashlar blocks,
almost completely preserved wooden formworks were re-
vealed, forming the foundation of the mole construction
(fig. 16). These caissons are 5.25 m long and 1.8 m wide and
filled with a rough conglomerate of quarry stones and mortar,
which most likely forms a certain type of hydraulic concrete
composition 152. According to the latest publications of the
excavation results that refer to studies on the analysis of the
mortar, its composition can allegedly be identified as the
pozzolanic mortar described by Vitruvius and Procopius 153.
Based on Brandon’s definition 154, it seems more likely that
this material no longer represents »Roman marine concrete«,
but rather follows Roman harbour construction techniques
by using a similar reacting aggregate. Consequently, it has
further to be examined whether the concrete mixture indeed
comprises Roman pozzolanic mortar or any other volcanic
ash or aggregate.
If the extent of 13 m in east-west direction and 7 m in
north-south direction mentioned by Karagöz is to be attrib-
uted to the size of the mole construction and its breakwater
substructure 155, then the dimension can be considered to be
relatively small. However, it seems to have been sufficient
for the demands of the harbour and to keep it operational
for centuries. Concerning the functional efficiency, waves
must have been able to break over the structure in order to
counteract the constant siltation process from the rivers by
creating controlled currents within the harbour basin. As such,
the breakwater must have protruded from the surface of the
sea. Level measures between +0.48 and +0.71 m indeed give
that impression. Consequently, similar to the breakwater at
150 Karagöz, Khrysopolis Liman 410; Karagöz, Chrysopolis 46.
151 Karagöz, Khrysopolis Liman 411; Karagöz, Chrysopolis 47; Karagöz, Khrysop-
olis – Scutari 3.
152 Karagöz, Khrysopolis Liman 413 fig. 13; Karagöz, Chrysopolis 47-48.
153 Vitruvius, De Architectura V 12. 3 (129 Rose / Müller-Strübing); Prokopios, De
Aedificiis I 11. 18-20 (IV 44 Haury
the Theodosian harbour, it can again be identified as of type
»Mound Breakwater«.
As for the dating of the harbour site, in contrast to the
large time span of the archaeological finds, the earliest har-
bour facilities are not to be dated before the Roman Imperial
period and most likely belong to Early Byzantine times, as
attested by the wooden piers 156. If any permanent harbour
structures of the Classical to Hellenistic periods ever existed,
then they must have been situated further inland. In contrast
to the general understanding of the harbour situation 157,
which is similar to the quayside in the western basin of the
Harbour of Theodosius (see above), any Roman coastal fa-
cilities must also be located slightly further southeast. These
may again have at least partly been re-used for any Early Byz-
antine harbour works. Unlike Procopius’ detailed description
of construction works at the harbour of Eutropiu Limen and
possibly also at Hiereia during the sixth century 158, no such
building activities at Chrysopolis are confirmed by any Early
Byzantine sources.
The building material, the marble column spolia of the
mole and the wooden piers located farther to the southeast
indicate that additional harbour infrastructures must have ex-
isted in the Early Byzantine period. This is supported by a large
number of ceramic artefacts such as oil lamps or Unguentaria,
all dating to the fifth to sixth centuries. In fact, these may
again belong to the extensive building activities under the
reign of Emperor Justinian I or his immediate predecessors.
Indeed, a considerable amount of African Red Slip Ware (ARS)
shows trading connections to North African markets follow-
ing its re-conquest and annexation in the 530s 159.
After the Early Byzantine period, the harbour obviously
suffered increasingly from constant siltation by river alluvium
and other environmental impacts, such as earthquakes, or
geopolitical events 160. This eventually led to the shift of the
harbour area farther north-west, but when did the relocation
of the harbour site and, accordingly, the erection of a new
harbour installation take place? While the pottery (particularly
the African Red Slip Ware) and the re-used spolia from the
preceding Early Byzantine harbour site provide a terminus
post quem of the sixth century, the construction of a large ec-
clesiastical complex on part of the supposed harbour basin in
the twelfth to thirteenth centuries may be taken as a terminus
ante quem 161. On the one hand, it is not just to assume that
it must have taken some time for the Early Byzantine harbour
site to become unusable for ships, thus making relocation
156 Karagöz, Khrysopolis Liman 408; Karagöz, Chrysopolis 49.
157 Belke, Gates 166, in this volume; Karagöz, Yapı 421-423.
158 Prokopios, De Aedificiis I 11. 16-23 (IV 43-45 Haury / Wirth); Belke, Gates 167.
170, in this volume; Hohlfelder, Building Harbours 368-370.
159 Karagöz, Khrysopolis Liman 414; Karagöz, Chrysopolis 44-45.